# AI Path Finder

> Free AI career consultation (AI role consultation / AI career counselling). A questionnaire matches software engineers, testers, DevOps and data engineers, freshers, MCA/BCA graduates and career changers from non-IT fields to the best-fit AI role. It explains why the role fits and gives a personalised, dated learning roadmap.

Last updated: 2026-10-28

## How the recommendation works
- About 20 questions in 5 sections: background and experience, skill levels (Python, SQL, cloud, AI/ML, maths), a 4-question reasoning check, working style and interests, goals and weekly study hours.
- Self-ratings are checked against the reasoning questions, and contradictory answers are corrected.
- Each of the 9 roles is scored with weighted, explainable signals. Users see a fit %, readiness %, skill gaps, transferable strengths and a roadmap that skips what they already know.
- Everything runs in the browser. Answers are not sent to a server.

## AI career paths
- [AI Foundations Path](https://homesh123.github.io/ai-career-advisor/roles/ai-foundations-beginner.html): Build the base, then land a junior AI / ML role. Best for: students, graduates and people new to coding. Time to job-ready: 4.8 months at 10 h/week. Coding: basic Python; maths: basic statistics.
- [ML Engineer](https://homesh123.github.io/ai-career-advisor/roles/ml-engineer.html): Build, train and ship ML models into real products. Best for: backend and data developers who enjoy data and maths. Time to job-ready: 3.9–5.3 months at 10 h/week. Coding: comfortable Python; maths: basic statistics.
- [Data Engineer (AI focus)](https://homesh123.github.io/ai-career-advisor/roles/ai-data-engineer.html): Build the data pipelines every AI system depends on. Best for: data engineers, BI analysts and SQL-heavy developers. Time to job-ready: 3–3.5 months at 10 h/week. Coding: comfortable Python; maths: very little.
- [AI Product Engineer](https://homesh123.github.io/ai-career-advisor/roles/ai-product-engineer.html): Ship AI features users love — chat, search, copilots. Best for: frontend, full-stack and mobile developers. Time to job-ready: 3–3.2 months at 10 h/week. Coding: basic Python; maths: very little.
- [MLOps / AI Platform Engineer](https://homesh123.github.io/ai-career-advisor/roles/mlops-engineer.html): Run AI reliably and cheaply at scale. Best for: DevOps, cloud, SRE and infrastructure engineers. Time to job-ready: 2.5–3.5 months at 10 h/week. Coding: comfortable Python; maths: very little.
- [LLM / Agent Engineer](https://homesh123.github.io/ai-career-advisor/roles/llm-agent-engineer.html): Build AI agents that plan, use tools and get work done. Best for: backend developers and automation / QA engineers. Time to job-ready: 2.5–3.2 months at 10 h/week. Coding: comfortable Python; maths: very little.
- [AI Research Engineer](https://homesh123.github.io/ai-career-advisor/roles/ai-research-engineer.html): Train, fine-tune and push the limits of models. Best for: people with strong maths who love deep technical work. Time to job-ready: 3.2–5.3 months at 10 h/week. Coding: advanced Python; maths: strong (linear algebra, calculus, probability).
- [AI Forward Deployed Engineer](https://homesh123.github.io/ai-career-advisor/roles/ai-forward-deployed-engineer.html): Deploy AI inside real businesses, hands-on with customers. Best for: senior engineers who enjoy clients and know an industry. Time to job-ready: 2.3–3.5 months at 10 h/week. Coding: comfortable Python; maths: very little.
- [AI Product Manager](https://homesh123.github.io/ai-career-advisor/roles/ai-product-manager.html): Decide what AI should build — and make it succeed. Best for: PMs, BAs, team leads and non-IT domain experts. Time to job-ready: 3–3.2 months at 10 h/week. Coding: not required (low-code is fine); maths: basic statistics.

## AI career guidance by background
- [Software testers and QA engineers](https://homesh123.github.io/ai-career-advisor/for/software-testers-qa.html): usually fits LLM / Agent Engineer, then AI Product Engineer or ML Engineer.
- [Java, .NET and backend developers](https://homesh123.github.io/ai-career-advisor/for/java-dotnet-backend-developers.html): usually fits LLM / Agent Engineer, then AI Product Engineer or ML Engineer.
- [Frontend and full-stack developers](https://homesh123.github.io/ai-career-advisor/for/frontend-full-stack-developers.html): usually fits AI Product Engineer, then AI Forward Deployed Engineer or AI Product Manager.
- [DevOps, cloud and SRE engineers](https://homesh123.github.io/ai-career-advisor/for/devops-cloud-engineers.html): usually fits MLOps / AI Platform Engineer, then AI Forward Deployed Engineer or ML Engineer.
- [Data engineers and BI analysts](https://homesh123.github.io/ai-career-advisor/for/data-engineers-analysts.html): usually fits Data Engineer (AI focus), then ML Engineer or AI Forward Deployed Engineer.
- [Freshers, students and MCA / BCA graduates](https://homesh123.github.io/ai-career-advisor/for/freshers-mca-bca-graduates.html): usually fits AI Foundations Path, then ML Engineer or AI Research Engineer.
- [Non-IT professionals (finance, healthcare, teaching, operations)](https://homesh123.github.io/ai-career-advisor/for/non-it-professionals.html): usually fits AI Product Manager, then AI Forward Deployed Engineer or AI Foundations Path.
- [Product managers, project managers and business analysts](https://homesh123.github.io/ai-career-advisor/for/product-managers-business-analysts.html): usually fits AI Product Manager, then AI Forward Deployed Engineer or AI Foundations Path.
- [Senior engineers, architects and tech leads (10+ years)](https://homesh123.github.io/ai-career-advisor/for/senior-engineers-tech-leads.html): usually fits AI Forward Deployed Engineer, then AI Product Manager or AI Product Engineer.

## Guides
- [How to choose the right AI role: a 5-step framework](https://homesh123.github.io/ai-career-advisor/guide/how-to-choose-an-ai-role.html): Pick the AI role that reuses most of what you already do, matches the kind of work you enjoy every day, and fits your honest comfort with coding and maths and the hours you can study. Then test it with one small project before you commit months to it. For most working professionals that means an adjacent move rather than starting over: backend to LLM / Agent Engineer, frontend to AI Product Engineer, DevOps to MLOps, data to AI Data Engineer, and PMs or domain experts to AI Product Manager.
- [How the AI career consultation works](https://homesh123.github.io/ai-career-advisor/guide/how-it-works.html): the questions, validation rules, scoring signals per role, and limitations.

## Learn: free AI courses (English and Hinglish)
- [Learn hub](https://homesh123.github.io/ai-career-advisor/learn.html): 9 courses, one per AI role, following each roadmap step by step. 54 lessons, each with a real-life analogy, a plain-language explanation, key terms, a worked example and a quick check. Every lesson is available in English and Hinglish (Hindi in Roman script with English technical terms).
- [What is AI? AI vs ML vs Deep Learning vs GenAI](https://homesh123.github.io/ai-career-advisor/learn/what-is-ai.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/what-is-ai.html): AI ⊃ Machine Learning ⊃ Deep Learning ⊃ Generative AI. Each is a more specific part of the one before.
- [Python basics: variables and data types](https://homesh123.github.io/ai-career-advisor/learn/py-basics.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/py-basics.html): A variable is a labelled box, and every value inside it has a type.
- [Python logic: if/else, loops and functions](https://homesh123.github.io/ai-career-advisor/learn/py-logic.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/py-logic.html): if = decide, loop = repeat, function = reuse.
- [Lists and dictionaries: storing many values](https://homesh123.github.io/ai-career-advisor/learn/py-collections.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/py-collections.html): List = ordered line of items. Dictionary = look up by name (key).
- [Git and GitHub: saving and sharing your work](https://homesh123.github.io/ai-career-advisor/learn/git-github.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/git-github.html): add → commit → push. Commit small and often, with clear messages.
- [Pandas: working with tables of data](https://homesh123.github.io/ai-career-advisor/learn/pandas.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/pandas.html): DataFrame = table. Most AI projects start with load → clean → groupby.
- [Averages and spread: mean, median, variance](https://homesh123.github.io/ai-career-advisor/learn/stats-center-spread.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/stats-center-spread.html): Use the median when there are outliers, and always look at the spread too.
- [Probability and distributions](https://homesh123.github.io/ai-career-advisor/learn/stats-probability.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/stats-probability.html): Model outputs are usually probabilities, and you pick the threshold.
- [Correlation vs causation, and sampling bias](https://homesh123.github.io/ai-career-advisor/learn/stats-correlation.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/stats-correlation.html): Correlation is not causation, and biased data gives biased AI.
- [SQL basics: asking a database questions](https://homesh123.github.io/ai-career-advisor/learn/sql-basics.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/sql-basics.html): SELECT what → FROM where → WHERE which rows → GROUP BY → ORDER BY.
- [SQL JOINs: combining tables](https://homesh123.github.io/ai-career-advisor/learn/sql-joins.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/sql-joins.html): INNER = only matches. LEFT = keep everything on the left.
- [How machines learn: features, labels and training](https://homesh123.github.io/ai-career-advisor/learn/ml-how-machines-learn.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-how-machines-learn.html): Features in → label out. Training = learning from examples with known answers.
- [Two kinds of prediction: regression and classification](https://homesh123.github.io/ai-career-advisor/learn/ml-regression-classification.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-regression-classification.html): Number → regression. Category → classification. Start simple.
- [Train/test split and overfitting](https://homesh123.github.io/ai-career-advisor/learn/ml-train-test.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-train-test.html): Never judge a model on data it trained on.
- [Measuring models: accuracy, precision and recall](https://homesh123.github.io/ai-career-advisor/learn/ml-metrics.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-metrics.html): Beat the baseline, and choose precision vs recall by the cost of each mistake.
- [What is an LLM? Tokens, context and temperature](https://homesh123.github.io/ai-career-advisor/learn/llm-what.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-what.html): LLM = super-powered next-word predictor. Mind tokens, context and temperature.
- [Prompting: giving clear instructions to an LLM](https://homesh123.github.io/ai-career-advisor/learn/llm-prompting.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-prompting.html): Role + context + task + format + examples. Then test and iterate.
- [Calling an LLM API: keys, requests and costs](https://homesh123.github.io/ai-career-advisor/learn/llm-api.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-api.html): Request in → response out, billed per token. Guard the key, and handle errors.
- [Building a portfolio project that gets interviews](https://homesh123.github.io/ai-career-advisor/learn/portfolio.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/portfolio.html): Real problem, end-to-end, evaluated, deployed and well explained.
- [Landing the AI role: resume, applications and interviews](https://homesh123.github.io/ai-career-advisor/learn/job-search.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/job-search.html): Action + result + number. Reframe your experience in AI terms. Apply steadily.
- [Feature engineering: turning raw data into useful inputs](https://homesh123.github.io/ai-career-advisor/learn/ml-features.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-features.html): Good features + domain knowledge > fancy algorithms. Watch out for leakage.
- [Decision trees, random forests and boosting](https://homesh123.github.io/ai-career-advisor/learn/ml-trees.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-trees.html): Tables → try boosted trees first.
- [Vectors and matrices: the language of AI](https://homesh123.github.io/ai-career-advisor/learn/math-vectors.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/math-vectors.html): Everything in AI becomes vectors, and similarity is just vector maths.
- [Gradients and gradient descent: how models improve](https://homesh123.github.io/ai-career-advisor/learn/math-gradients.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/math-gradients.html): Gradient = which way is downhill. Learning = many small steps down the loss hill.
- [Neural networks: layers that learn](https://homesh123.github.io/ai-career-advisor/learn/dl-neural-networks.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/dl-neural-networks.html): Layers of weighted sums plus activations, trained by backprop.
- [Transformers and attention: how LLMs read](https://homesh123.github.io/ai-career-advisor/learn/dl-transformers.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/dl-transformers.html): Tokens → vectors → many attention layers → next-token prediction.
- [Training vs inference: how ML runs in production](https://homesh123.github.io/ai-career-advisor/learn/ml-training-vs-inference.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ml-training-vs-inference.html): Training creates the model. Inference serves it, fast, cheaply and reliably.
- [Serving a model as an API with FastAPI](https://homesh123.github.io/ai-career-advisor/learn/api-serving.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/api-serving.html): Model → API endpoint → any app can use it. Validate inputs and log everything.
- [Docker: packaging your app so it runs anywhere](https://homesh123.github.io/ai-career-advisor/learn/docker.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/docker.html): Dockerfile (recipe) → image (package) → container (running app).
- [MLOps: experiment tracking and model registries](https://homesh123.github.io/ai-career-advisor/learn/mlops-tracking.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/mlops-tracking.html): Log every run, register every shipped model, and always be able to roll back.
- [Monitoring models: drift and silent failures](https://homesh123.github.io/ai-career-advisor/learn/mlops-monitoring.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/mlops-monitoring.html): Models decay silently. Monitor inputs, outputs and outcomes, with alerts.
- [Data modelling: designing tables for analysis](https://homesh123.github.io/ai-career-advisor/learn/data-modeling.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/data-modeling.html): Facts = what happened (numbers). Dimensions = descriptions.
- [Data pipelines: ETL, Airflow, dbt and Spark](https://homesh123.github.io/ai-career-advisor/learn/data-pipelines.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/data-pipelines.html): Extract → transform → load, orchestrated, tested, re-runnable and monitored.
- [Data quality: catching bad data early](https://homesh123.github.io/ai-career-advisor/learn/data-quality.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/data-quality.html): Check completeness, validity, uniqueness, consistency and freshness, automatically.
- [Embeddings and vector databases: searching by meaning](https://homesh123.github.io/ai-career-advisor/learn/embeddings.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/embeddings.html): Embeddings = meaning as numbers. A vector DB finds the nearest meanings fast.
- [RAG: letting an LLM answer from your documents](https://homesh123.github.io/ai-career-advisor/learn/rag.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/rag.html): RAG = search your docs, then answer from them with citations.
- [Streaming data: Kafka and real-time pipelines](https://homesh123.github.io/ai-career-advisor/learn/streaming.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/streaming.html): Batch = periodic, simpler. Streaming = real-time, more complex. Use it when seconds matter.
- [Evaluating AI: how to know it actually works](https://homesh123.github.io/ai-career-advisor/learn/llm-evals.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-evals.html): No evals, no confidence. Measure every change.
- [AI safety in products: hallucinations, prompt injection and guardrails](https://homesh123.github.io/ai-career-advisor/learn/llm-safety.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-safety.html): Treat model output and documents as untrusted. Limit what tools can do.
- [UX for AI features: designing for uncertainty](https://homesh123.github.io/ai-career-advisor/learn/ai-ux.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ai-ux.html): Stream it, source it, let users edit it, and learn from feedback.
- [Tool use (function calling): letting an LLM take actions](https://homesh123.github.io/ai-career-advisor/learn/llm-tool-use.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-tool-use.html): The model asks, your code acts. Validate everything.
- [Serving LLMs efficiently: GPUs, batching and cost](https://homesh123.github.io/ai-career-advisor/learn/gpu-serving.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/gpu-serving.html): Batch requests, measure latency and cost, and quantise and cache where you can.
- [Kubernetes: running containers at scale](https://homesh123.github.io/ai-career-advisor/learn/kubernetes.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/kubernetes.html): Declare the desired state, and Kubernetes keeps it true: restarts, scaling, updates.
- [AI agents: plan, act, observe, repeat](https://homesh123.github.io/ai-career-advisor/learn/llm-agents.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-agents.html): Agent = LLM + tools + loop + memory + guardrails.
- [MCP: a universal plug for AI tools](https://homesh123.github.io/ai-career-advisor/learn/llm-mcp.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/llm-mcp.html): MCP = one standard plug between AI apps and your tools and data.
- [Fine-tuning and LoRA: specialising a model](https://homesh123.github.io/ai-career-advisor/learn/dl-fine-tuning.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/dl-fine-tuning.html): Prompting → RAG → fine-tuning, in that order of cost.
- [Reading and reproducing research papers](https://homesh123.github.io/ai-career-advisor/learn/research-papers.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/research-papers.html): Skim first, then go deep, then reproduce. Reproduction is proof.
- [Enterprise AI: integration, security and compliance](https://homesh123.github.io/ai-career-advisor/learn/security-compliance.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/security-compliance.html): Enforce access, mask PII, log everything, and involve security early.
- [Problem framing: should AI solve this, and how?](https://homesh123.github.io/ai-career-advisor/learn/problem-framing.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/problem-framing.html): Problem → metric → data → simplest solution. Diagnose before you prescribe.
- [Explaining AI to non-technical people](https://homesh123.github.io/ai-career-advisor/learn/stakeholder-communication.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/stakeholder-communication.html): Outcome → evidence → risk → ask. Analogies over jargon. Demos over slides.
- [AI literacy for decisions: build vs buy, cost and limits](https://homesh123.github.io/ai-career-advisor/learn/build-vs-buy.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/build-vs-buy.html): Buy → API → RAG → fine-tune → build. Climb only when needed.
- [Prototyping AI ideas with low-code tools](https://homesh123.github.io/ai-career-advisor/learn/no-code-prototyping.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/no-code-prototyping.html): Prototype in days, test with real users, then write the spec.
- [Metrics and A/B testing for AI features](https://homesh123.github.io/ai-career-advisor/learn/ab-testing.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ab-testing.html): Define success and guardrails first, then randomise and wait for significance.
- [Writing an AI product spec (PRD) with an eval plan](https://homesh123.github.io/ai-career-advisor/learn/ai-prd.html) · [Hinglish](https://homesh123.github.io/ai-career-advisor/learn/hi/ai-prd.html): Problem, metric, scope, UX for failure, eval plan, risks, launch bar.

## Interview practice
- [Interview practice with scoring](https://homesh123.github.io/ai-career-advisor/interview.html): role-specific questions (multiple-choice scored automatically, open questions self-checked against model answers). The report gives a score per skill area, explains each mistake and links the lessons for every gap.
- [AI Foundations Path interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ai-foundations-beginner.html): 38 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 21 quiz questions. Typical rounds: HR or recruiter screen (15 to 30 min); Online test; Technical interview; Managerial or HR round.
- [ML Engineer interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ml-engineer.html): 53 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 27 quiz questions. Typical rounds: Recruiter screen; Coding round; ML fundamentals; ML system design; Behavioural and project deep dive.
- [Data Engineer (AI focus) interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ai-data-engineer.html): 42 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 26 quiz questions. Typical rounds: Recruiter screen; SQL round; Python or Spark round; Pipeline and data modelling design; Behavioural.
- [AI Product Engineer interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ai-product-engineer.html): 43 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 24 quiz questions. Typical rounds: Recruiter screen; Coding round; AI feature design; System design; Craft deep dive on an AI feature you built, plus behavioural.
- [MLOps / AI Platform Engineer interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/mlops-engineer.html): 35 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 16 quiz questions. Typical rounds: Recruiter screen; Coding or scripting round; MLOps concepts; System design; Incident and behavioural round.
- [LLM / Agent Engineer interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/llm-agent-engineer.html): 44 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 22 quiz questions. Typical rounds: Recruiter screen; Coding round; RAG and evals deep dive; System design; Project deep dive and behavioural.
- [AI Research Engineer interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ai-research-engineer.html): 34 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 18 quiz questions. Typical rounds: Recruiter screen; ML coding; Math and fundamentals; Research discussion; Behavioural and team fit.
- [AI Forward Deployed Engineer interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ai-forward-deployed-engineer.html): 37 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 23 quiz questions. Typical rounds: Recruiter screen; Coding round; Decomposition case (the hardest round); Technical depth; Customer and behavioural round.
- [AI Product Manager interview questions and answers](https://homesh123.github.io/ai-career-advisor/interview/ai-product-manager.html): 39 commonly asked questions in interview order (introduction, fundamentals, role skills, design, behaviour, salary) with easy model answers, plus 24 quiz questions. Typical rounds: Recruiter screen; Product sense; Analytical and metrics; AI depth; Behavioural and leadership.

## Resume builder and ATS-friendly samples
- [Resume builder](https://homesh123.github.io/ai-career-advisor/resume-builder.html): upload a resume (PDF, Word .docx or text) or start from a role sample, pick the target AI role and add missing ATS keywords, choose one of three single-column templates, edit with a live preview and a "human tone" check, and download as PDF or Word. The file is read in the browser and never uploaded. It can also pre-fill the career questionnaire.
- [All resume samples](https://homesh123.github.io/ai-career-advisor/resumes/): editable in the browser, downloadable as Word, PDF or text, with a live ATS keyword checklist.
- [AI Foundations Path resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ai-foundations-beginner.html): keywords Python, SQL, Machine Learning, scikit-learn, Pandas, NumPy, Statistics, Git, Classification, Regression, LLM, Streamlit, Data Cleaning, Feature Engineering
- [ML Engineer resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ml-engineer.html): keywords Python, PyTorch, scikit-learn, XGBoost, Feature Engineering, Model Deployment, FastAPI, Docker, MLflow, MLOps, SQL, AWS, A/B Testing, Model Monitoring
- [Data Engineer (AI focus) resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ai-data-engineer.html): keywords SQL, Python, Apache Airflow, dbt, Apache Spark, Kafka, Data Modeling, ETL, Data Quality, Snowflake, Vector Database, Embeddings, RAG, Data Pipelines
- [AI Product Engineer resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ai-product-engineer.html): keywords TypeScript, React, Next.js, Node.js, Python, LLM, RAG, Vector Database, Prompt Engineering, LLM Evaluation, Streaming, Vercel AI SDK, A/B Testing, FastAPI
- [MLOps / AI Platform Engineer resume sample](https://homesh123.github.io/ai-career-advisor/resumes/mlops-engineer.html): keywords Kubernetes, Docker, Helm, Terraform, CI/CD, MLOps, MLflow, vLLM, GPU, Prometheus, Grafana, Python, AWS, Model Serving, Observability
- [LLM / Agent Engineer resume sample](https://homesh123.github.io/ai-career-advisor/resumes/llm-agent-engineer.html): keywords Python, LLM, AI Agents, LangGraph, Tool Calling, MCP, RAG, FastAPI, Pydantic, Evaluation, Prompt Engineering, Guardrails, Observability, Java
- [AI Research Engineer resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ai-research-engineer.html): keywords PyTorch, Deep Learning, Transformers, Fine tuning, LoRA, PEFT, Hugging Face, Distributed Training, Python, Linear Algebra, Experiment Tracking, NLP, Evaluation, CUDA
- [AI Forward Deployed Engineer resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ai-forward-deployed-engineer.html): keywords Python, LLM, RAG, Solution Architecture, Stakeholder Management, Client facing, API Integration, SSO, Security, Compliance, HIPAA, Prototyping, Requirements Gathering, Cloud
- [AI Product Manager resume sample](https://homesh123.github.io/ai-career-advisor/resumes/ai-product-manager.html): keywords Product Management, AI Product Strategy, LLM, Generative AI, PRD, Evaluation, A/B Testing, SQL, Stakeholder Management, Roadmap, User Research, Metrics, Responsible AI, Prototyping

## AI dictionary (A to Z)
- [AI dictionary](https://homesh123.github.io/ai-career-advisor/glossary/): 138 AI terms explained in English and Hinglish, each with an example and a lesson link.
- [What is A/B test?](https://homesh123.github.io/ai-career-advisor/glossary/a-b-test.html): Randomly showing two versions to different users to see which one really works better. Simply put: Like a sweet shop selling the same laddoo in two box designs for a week, then keeping the box that sold more.
- [What is Ablation study?](https://homesh123.github.io/ai-career-advisor/glossary/ablation-study.html): Removing one part of a system at a time to see how much each part really helps. Simply put: Like cooking dal three times, each time leaving out one spice, to find out which spice really makes it tasty.
- [What is Accuracy?](https://homesh123.github.io/ai-career-advisor/glossary/accuracy.html): The share of all predictions that were correct. Simply put: Like your exam score: 45 out of 50 questions right means 90%. But if almost every answer is 'No', always writing 'No' also scores high, so accuracy alone can fool you.
- [What is Activation function?](https://homesh123.github.io/ai-career-advisor/glossary/activation-function.html): A small maths function inside each neuron that lets neural networks learn curves, not just straight lines. Simply put: Like a doorkeeper for each brain cell: weak signals are stopped at the door, strong ones are let through. That on and off choice lets the network learn bends and curves.
- [What is Agent?](https://homesh123.github.io/ai-career-advisor/glossary/agent.html): An LLM that works in a loop: plans a step, uses a tool, looks at the result, and decides the next step until the goal is done. Simply put: Like a smart assistant you give a goal, not steps: 'book me a cheap train to Pune on Friday'. It searches, compares, checks your calendar and comes back to you before paying.
- [What is AI?](https://homesh123.github.io/ai-career-advisor/glossary/ai.html): Computer systems that do things we'd normally call smart: understand language, recognise images, make decisions.
- [What is AI engineer?](https://homesh123.github.io/ai-career-advisor/glossary/ai-engineer.html): A software engineer who builds products and systems using AI models, especially LLMs. Simply put: Like an electrician who doesn't make the power plant (the AI model) but wires it safely into your house so the lights and fans actually work.
- [What is Alignment?](https://homesh123.github.io/ai-career-advisor/glossary/alignment.html): Making AI models behave the way people intend: helpful, honest and safe. Simply put: Like training a very capable new employee in the company's values, so they are helpful to customers, honest, and refuse harmful requests.
- [What is Annotation?](https://homesh123.github.io/ai-career-advisor/glossary/annotation.html): Adding the correct answers (labels) to data so a model can learn from it. Simply put: Like a teacher writing the right answer next to each practice question, so the student can learn by checking.
- [What is API?](https://homesh123.github.io/ai-career-advisor/glossary/api.html): A defined way for one program to ask another program for data or an action. Simply put: Like a waiter in a restaurant: you don't go into the kitchen, you tell the waiter what you want and the waiter brings it back.
- [What is Attention?](https://homesh123.github.io/ai-career-advisor/glossary/attention.html): The mechanism in transformers that lets each word look at every other word to understand its meaning in context. Simply put: Like reading a sentence with a highlighter: to understand the word 'it', your eyes jump back to the words that tell you what 'it' is.
- [What is AUC?](https://homesh123.github.io/ai-career-advisor/glossary/auc.html): A single number summarising how well a model separates positives from negatives across all thresholds. Simply put: Like checking how well a guard can tell thieves from shoppers: pick one thief and one shopper at random, and AUC is how often the guard rates the thief as more suspicious.
- [What is Autoscaling?](https://homesh123.github.io/ai-career-advisor/glossary/autoscaling.html): Automatically adding or removing servers as traffic goes up or down. Simply put: Like a supermarket opening more billing counters when the queue gets long and closing them when it's quiet.
- [What is Backpropagation?](https://homesh123.github.io/ai-career-advisor/glossary/backpropagation.html): The method neural networks use to work out how much each weight contributed to the error, so it can be adjusted. Simply put: Like a cricket coach watching a lost match on replay and working out how much each player's mistake cost, so each one knows what to fix.
- [What is Baseline?](https://homesh123.github.io/ai-career-advisor/glossary/baseline.html): A simple reference result that any new model must beat to be worth using. Simply put: Like checking a new diet against simply eating normally: if the fancy plan doesn't beat the simple one, it isn't worth it.
- [What is Batch inference?](https://homesh123.github.io/ai-career-advisor/glossary/batch-inference.html): Running a model on many records at once on a schedule, instead of one request at a time. Simply put: Like a teacher correcting all 40 notebooks in one go at night, instead of correcting each one the moment a student hands it in.
- [What is Batch size?](https://homesh123.github.io/ai-career-advisor/glossary/batch-size.html): How many examples a model processes together in one training step. Simply put: Like a teacher correcting 32 copies, then giving the class one round of feedback, instead of giving feedback after every single copy.
- [What is Benchmark?](https://homesh123.github.io/ai-career-advisor/glossary/benchmark.html): A standard test set used to compare models fairly. Simply put: Like a board exam: every student gets the same paper, so the scores can be compared fairly.
- [What is Bias?](https://homesh123.github.io/ai-career-advisor/glossary/bias.html): When a model's results are systematically unfair or skewed, often because of skewed training data. Simply put: Like a cricket selector who has only ever watched players from one city: without meaning to, they keep picking from that city.
- [What is Bias variance trade off?](https://homesh123.github.io/ai-career-advisor/glossary/bias-variance-trade-off.html): The balance between a model that's too simple (misses patterns) and too complex (memorises noise). Simply put: Like studying for an exam: only reading the summary misses things (too simple), but mugging up last year's paper word by word fails on new questions (too complex). The sweet spot is understanding the ideas.
- [What is BM25?](https://homesh123.github.io/ai-career-advisor/glossary/bm25.html): A classic search method that ranks documents by matching words, used with vector search in hybrid search. Simply put: Like the index at the back of a textbook: it finds pages that contain the exact word you look up.
- [What is Chain of thought?](https://homesh123.github.io/ai-career-advisor/glossary/chain-of-thought.html): Asking a model to reason step by step before giving the final answer. Simply put: Like a maths teacher saying 'show your working': writing each step makes careless mistakes less likely.
- [What is Chatbot?](https://homesh123.github.io/ai-career-advisor/glossary/chatbot.html): A program you talk with in natural language; modern chatbots are usually powered by LLMs.
- [What is Chunking?](https://homesh123.github.io/ai-career-advisor/glossary/chunking.html): Splitting documents into smaller pieces before embedding them for search and RAG. Simply put: Like cutting a big pizza into slices so you can pick up exactly the piece you want.
- [What is CI/CD?](https://homesh123.github.io/ai-career-advisor/glossary/ci-cd.html): Automatically testing and releasing code (and models) every time something changes. Simply put: Like a factory quality check on every product before it leaves: each change is tested automatically, and only good ones are shipped.
- [What is Citations?](https://homesh123.github.io/ai-career-advisor/glossary/citations.html): Showing which document or passage an AI answer came from, so people can check it. Simply put: Like the footnotes in a school project that say where each fact came from, so the teacher can check it.
- [What is Classification?](https://homesh123.github.io/ai-career-advisor/glossary/classification.html): Predicting a category, like spam or not spam, or which department a ticket belongs to. Simply put: Like sorting the post into boxes: bills, letters, ads. Each item goes into one category.
- [What is Clustering?](https://homesh123.github.io/ai-career-advisor/glossary/clustering.html): Grouping similar items together without any labels. Simply put: Like tipping a box of mixed buttons on the table and grouping them by look, without anyone telling you the groups.
- [What is Computer vision?](https://homesh123.github.io/ai-career-advisor/glossary/computer-vision.html): AI that understands images and video: recognising objects, faces, text and scenes. Simply put: Giving a computer eyes: like how you glance at a photo and instantly see a dog, a car and a street.
- [What is Concept drift?](https://homesh123.github.io/ai-career-advisor/glossary/concept-drift.html): When the relationship between inputs and the right answer changes over time, so a trained model gets worse. Simply put: Like a shopkeeper whose old rule 'people buy umbrellas in June' stops working because the monsoon now starts in July.
- [What is Confusion matrix?](https://homesh123.github.io/ai-career-advisor/glossary/confusion-matrix.html): A 2×2 (or bigger) table showing correct and wrong predictions for each class. Simply put: Like a report card for a security guard: thieves caught, thieves missed, innocent people stopped, and innocent people let through.
- [What is Context window?](https://homesh123.github.io/ai-career-advisor/glossary/context-window.html): The maximum amount of text (in tokens) an LLM can consider at once, including its answer. Simply put: Like the size of a desk: everything the model can look at right now must fit on it. Older papers fall off the edge.
- [What is Cosine similarity?](https://homesh123.github.io/ai-career-advisor/glossary/cosine-similarity.html): A measure of how similar two vectors point, used to compare the meaning of embeddings. Simply put: Like checking if two people are pointing in the same direction: it doesn't matter how long their arms are, only where they point.
- [What is Cross entropy?](https://homesh123.github.io/ai-career-advisor/glossary/cross-entropy.html): The standard loss function for classification: it punishes confident wrong answers heavily. Simply put: Like a quiz that gives a small minus for an unsure wrong answer but a huge minus when you were 100% sure and still wrong.
- [What is Cross validation?](https://homesh123.github.io/ai-career-advisor/glossary/cross-validation.html): Rotating which part of the data is used for validation to get a more reliable score. Simply put: Like taking five different mock tests instead of one, so a single lucky or unlucky paper doesn't fool you.
- [What is Data drift?](https://homesh123.github.io/ai-career-advisor/glossary/data-drift.html): When the data a live model receives starts to look different from its training data. Simply put: Like a weather app trained on Delhi now being used in Chennai: the inputs look different from what it learned on.
- [What is Data leakage?](https://homesh123.github.io/ai-career-advisor/glossary/data-leakage.html): When information that won't exist at prediction time sneaks into training, making test scores falsely high. Simply put: Like a student who saw the answer key before the exam: the score looks amazing, but it says nothing about what they really know.
- [What is Data pipeline?](https://homesh123.github.io/ai-career-advisor/glossary/data-pipeline.html): Automated steps that move data from sources, clean and transform it, and deliver it for reports or models. Simply put: Like a water supply system: water is collected, filtered and piped to every tap, automatically every day.
- [What is Data warehouse?](https://homesh123.github.io/ai-career-advisor/glossary/data-warehouse.html): A database designed for analysing large amounts of historical data with SQL. Simply put: Like a big, well organised godown of a company's history, arranged so you can quickly answer questions like 'sales by city last year'.
- [What is Dataset?](https://homesh123.github.io/ai-career-advisor/glossary/dataset.html): A collection of examples used to train, validate or test a model. Simply put: Like a big recipe book of past examples that the model studies from.
- [What is Decision tree?](https://homesh123.github.io/ai-career-advisor/glossary/decision-tree.html): A model that predicts by asking a series of yes or no questions about the input. Simply put: Like the game '20 questions': each yes or no question narrows things down until you reach the answer.
- [What is Deep learning?](https://homesh123.github.io/ai-career-advisor/glossary/deep-learning.html): Machine learning with neural networks of many layers, great for images, speech and text. Simply put: Like learning to recognise a face in stages: first edges, then eyes and nose, then the whole face. Each layer builds on the one before.
- [What is Deployment?](https://homesh123.github.io/ai-career-advisor/glossary/deployment.html): Making a model available for real use, usually as an API or a scheduled job. Simply put: Like moving a dish from your home kitchen to a restaurant menu: now real customers can order it any time.
- [What is Diffusion model?](https://homesh123.github.io/ai-career-advisor/glossary/diffusion-model.html): A generative model that creates images by starting from noise and removing it step by step. Simply put: Like a sculptor starting with a rough block and chipping away a little at a time until a statue appears. Here the block is random noise.
- [What is Docker?](https://homesh123.github.io/ai-career-advisor/glossary/docker.html): A tool that packages an app with everything it needs so it runs the same everywhere. Simply put: Like a tiffin box that carries the whole meal with its own spoon and chutney, so it tastes the same in any office.
- [What is DPO?](https://homesh123.github.io/ai-career-advisor/glossary/dpo.html): A way to fine tune LLMs on pairs of preferred and rejected answers, simpler than RLHF. Simply put: Like showing a trainee two replies and saying 'this one is better, that one is worse', again and again, until they naturally write the better kind.
- [What is Dropout?](https://homesh123.github.io/ai-career-advisor/glossary/dropout.html): Randomly switching off some neurons during training so the network doesn't rely on any single one. Simply put: Like a team practising with random players sitting out, so everyone learns every role and nobody depends on one star.
- [What is Early stopping?](https://homesh123.github.io/ai-career-advisor/glossary/early-stopping.html): Stopping training when the validation score stops improving, to avoid overfitting. Simply put: Like taking the cake out of the oven when it's perfectly done, instead of baking longer and burning it.
- [What is Embedding?](https://homesh123.github.io/ai-career-advisor/glossary/embedding.html): A list of numbers that captures the meaning of text, images or items, so similar meanings are close together. Simply put: Like giving every word a location on a map, where words with similar meaning live in the same neighbourhood.
- [What is Epoch?](https://homesh123.github.io/ai-career-advisor/glossary/epoch.html): One full pass of the training algorithm over the entire training dataset. Simply put: Like reading your whole textbook once from cover to cover. Three epochs means three full readings.
- [What is ETL and ELT?](https://homesh123.github.io/ai-career-advisor/glossary/etl-and-elt.html): Ways to move data: transform before loading (ETL) or load raw data then transform inside the warehouse (ELT). Simply put: Like groceries: ETL is washing and cutting vegetables before putting them in the fridge; ELT is putting everything in the fridge first and cutting what you need later.
- [What is Evals?](https://homesh123.github.io/ai-career-advisor/glossary/evals.html): A repeatable set of test cases and scoring rules that measures whether an AI system works well. Simply put: Like a fixed set of practice tests you rerun after every change, to see if the system got better or worse.
- [What is Explainability?](https://homesh123.github.io/ai-career-advisor/glossary/explainability.html): Understanding and showing why a model made a particular prediction. Simply put: Like a doctor who doesn't just say 'take rest', but explains 'your fever and these test results are why'.
- [What is F1 score?](https://homesh123.github.io/ai-career-advisor/glossary/f1-score.html): One number that balances precision and recall. Simply put: Like judging a fisherman on two things at once: how many of the fish in the lake he caught, and how much of his catch was actually fish and not old shoes.
- [What is Feature?](https://homesh123.github.io/ai-career-advisor/glossary/feature.html): One input the model uses to make a prediction, like age, income or number of logins. Simply put: Like the clues a property dealer uses to guess a flat's price: size, floor, area and distance from the station.
- [What is Feature engineering?](https://homesh123.github.io/ai-career-advisor/glossary/feature-engineering.html): Turning raw data into useful model inputs, often using domain knowledge. Simply put: Like a cook turning raw ingredients into ready masalas: the model gets inputs that are much more useful than the raw data.
- [What is Feature store?](https://homesh123.github.io/ai-career-advisor/glossary/feature-store.html): A shared system that stores ready-made features for both training and live predictions. Simply put: Like a shared pantry of ready made masalas, so every cook uses exactly the same mix in training and in the real kitchen.
- [What is Few shot prompting?](https://homesh123.github.io/ai-career-advisor/glossary/few-shot-prompting.html): Showing an LLM a few examples of the task inside the prompt so it copies the pattern. Simply put: Like showing a new helper three folded shirts before asking them to fold the rest.
- [What is Fine tuning?](https://homesh123.github.io/ai-career-advisor/glossary/fine-tuning.html): Further training a pretrained model on your own examples to specialise its behaviour. Simply put: Like a general doctor doing a short extra course to become a skin specialist: the basics stay, the focus sharpens.
- [What is Foundation model?](https://homesh123.github.io/ai-career-advisor/glossary/foundation-model.html): A large model trained on broad data that can be adapted to many tasks. Simply put: Like a graduate with broad general knowledge who can be trained quickly for many different jobs.
- [What is GenAI?](https://homesh123.github.io/ai-career-advisor/glossary/genai.html): AI that creates new content: text, images, audio, video or code. Simply put: Like a helper who doesn't just find an existing answer but writes, draws or composes something new for you.
- [What is GPU?](https://homesh123.github.io/ai-career-advisor/glossary/gpu.html): A chip that does thousands of maths operations in parallel, which makes training and running AI fast. Simply put: Like 1,000 students each solving one small sum at the same time, instead of one genius solving 1,000 sums one by one.
- [What is Gradient?](https://homesh123.github.io/ai-career-advisor/glossary/gradient.html): The direction and steepness in which the error grows; models step the opposite way to improve. Simply put: Like feeling which way the ground slopes under your feet: it tells you which direction makes the error go up, so you step the other way.
- [What is Gradient boosting?](https://homesh123.github.io/ai-career-advisor/glossary/gradient-boosting.html): Building many small decision trees one after another, each fixing the previous ones' mistakes. Simply put: Like a student who, after each mock test, studies only the questions they got wrong. Many small rounds add up to a strong result.
- [What is Gradient descent?](https://homesh123.github.io/ai-career-advisor/glossary/gradient-descent.html): The training method that repeatedly adjusts weights in small steps to reduce error. Simply put: Like walking downhill in fog: you can't see the bottom, so you keep taking small steps where the ground slopes down.
- [What is Grounding?](https://homesh123.github.io/ai-career-advisor/glossary/grounding.html): Making an LLM base its answer on provided sources or tool results, not just its memory. Simply put: Like an open book exam: answer from the book in front of you, not from fuzzy memory.
- [What is Guardrail metric?](https://homesh123.github.io/ai-career-advisor/glossary/guardrail-metric.html): A metric that must not get worse while you improve the main metric. Simply put: Like dieting to lose weight while making sure your energy levels don't crash: one number improves, another must not get worse.
- [What is Guardrails?](https://homesh123.github.io/ai-career-advisor/glossary/guardrails.html): Rules and checks around an AI system that block unsafe, wrong or costly behaviour. Simply put: Like the railings on a mountain road: you can still drive freely, but they stop you going over the edge.
- [What is Hallucination?](https://homesh123.github.io/ai-career-advisor/glossary/hallucination.html): When an AI confidently states something false or made up. Simply put: Like a student who doesn't know the answer but writes a confident, made up one to fill the page.
- [What is Human in the loop?](https://homesh123.github.io/ai-career-advisor/glossary/human-in-the-loop.html): Keeping a person involved to review or approve AI output before it takes effect. Simply put: Like a junior writing a letter and the manager reading and signing it before it is sent.
- [What is Hybrid search?](https://homesh123.github.io/ai-career-advisor/glossary/hybrid-search.html): Combining keyword search and vector search to get the best of both. Simply put: Like asking both a librarian who knows exact book titles and one who understands what you mean, then combining their picks.
- [What is Hyperparameter?](https://homesh123.github.io/ai-career-advisor/glossary/hyperparameter.html): A setting you choose before training, like learning rate or tree depth, rather than something the model learns. Simply put: Like the oven temperature and timer you set before baking: the cake doesn't choose them, you do.
- [What is Idempotency?](https://homesh123.github.io/ai-career-advisor/glossary/idempotency.html): A job is idempotent if running it twice gives the same result as running it once. Simply put: Like pressing a lift button twice: the lift still comes once.
- [What is Inference?](https://homesh123.github.io/ai-career-advisor/glossary/inference.html): Using a trained model to make predictions on new data. Simply put: Training is studying; inference is sitting the exam: using what was learned to answer a new question.
- [What is Instruction tuning?](https://homesh123.github.io/ai-career-advisor/glossary/instruction-tuning.html): Fine tuning a base model on instruction and answer examples so it follows requests helpfully. Simply put: Like teaching a well read person good manners at a help desk: listen to the request, then answer it clearly.
- [What is K means?](https://homesh123.github.io/ai-career-advisor/glossary/k-means.html): A clustering method that groups items around K centre points. Simply put: Like placing 4 ice cream carts in a park so each visitor walks to the nearest one, then moving the carts until each sits in the middle of its crowd.
- [What is Kubernetes?](https://homesh123.github.io/ai-career-advisor/glossary/kubernetes.html): A system that runs and manages many containers across many servers: starting, scaling and healing them. Simply put: Like a building manager for containers: decides which flat each one lives in, adds more when it gets busy and replaces any that break.
- [What is KV cache?](https://homesh123.github.io/ai-career-advisor/glossary/kv-cache.html): Memory that stores past attention keys and values so an LLM doesn't recompute them for every new token. Simply put: Like keeping notes while reading a long story, so for each new page you don't have to reread the whole book.
- [What is Label?](https://homesh123.github.io/ai-career-advisor/glossary/label.html): The correct answer for an example, which a supervised model learns to predict. Simply put: The answer written on the back of a flashcard.
- [What is Latency?](https://homesh123.github.io/ai-career-advisor/glossary/latency.html): How long a request takes from asking to getting the answer. Simply put: Like the time between ordering a chai and getting it in your hand.
- [What is Learning rate?](https://homesh123.github.io/ai-career-advisor/glossary/learning-rate.html): How big a step the model takes when updating its weights during training. Simply put: Like the size of your steps when looking for something in the dark: too big and you keep overshooting, too small and it takes forever.
- [What is LLM?](https://homesh123.github.io/ai-career-advisor/glossary/llm.html): A very large neural network trained on huge amounts of text to predict the next word, which lets it write, answer and reason. Simply put: Like a super powered autocomplete that has read a huge library, so it can continue any text in a sensible way: answers, emails, code.
- [What is LLM as judge?](https://homesh123.github.io/ai-career-advisor/glossary/llm-as-judge.html): Using an LLM with a rubric to grade another model's answers. Simply put: Like a senior teacher using a marking scheme to grade answer sheets written by another teacher's class.
- [What is LoRA?](https://homesh123.github.io/ai-career-advisor/glossary/lora.html): A cheap fine tuning method that trains small add-on matrices instead of the whole model. Simply put: Like adding a small clip on lens to your phone camera instead of buying a whole new camera.
- [What is Loss function?](https://homesh123.github.io/ai-career-advisor/glossary/loss-function.html): A formula that measures how wrong the model's predictions are; training tries to make it small. Simply put: Like the 'how far off was I?' score after each guess. Training keeps adjusting to make that score smaller.
- [What is Lost in the middle?](https://homesh123.github.io/ai-career-advisor/glossary/lost-in-the-middle.html): LLMs tend to use information at the start and end of a long prompt better than information in the middle. Simply put: Like remembering the first and last songs of a long playlist but not the ones in the middle.
- [What is Machine learning?](https://homesh123.github.io/ai-career-advisor/glossary/machine-learning.html): The part of AI where computers learn patterns from examples instead of following hand written rules. Simply put: Like learning to spot ripe mangoes by looking at hundreds of them, instead of someone writing you a rule book.
- [What is MCP?](https://homesh123.github.io/ai-career-advisor/glossary/mcp.html): An open standard for connecting AI apps to tools and data, like a universal plug. Simply put: Like a universal charger port: any app that supports it can plug into any tool that supports it.
- [What is Metric?](https://homesh123.github.io/ai-career-advisor/glossary/metric.html): A number you track to judge whether something is working. Simply put: Like the speedometer and fuel gauge in a car: numbers you watch to know how things are going.
- [What is MLOps?](https://homesh123.github.io/ai-career-advisor/glossary/mlops.html): The practices and tools for putting ML models into production and keeping them healthy. Simply put: Like running a restaurant kitchen, not just inventing a recipe: steady supplies, quality checks and fixing things fast when they go wrong.
- [What is Model?](https://homesh123.github.io/ai-career-advisor/glossary/model.html): The learned function that turns inputs into predictions or generated output. Simply put: Like a recipe the computer worked out by itself from examples: put inputs in, get a prediction out.
- [What is Model registry?](https://homesh123.github.io/ai-career-advisor/glossary/model-registry.html): A catalogue of trained model versions with their metrics and stage (testing, production). Simply put: Like a library catalogue for models: which version exists, how good it was, and which one is on the shelf right now.
- [What is Monitoring?](https://homesh123.github.io/ai-career-advisor/glossary/monitoring.html): Continuously checking a live system's health, data and results, with alerts when something goes wrong. Simply put: Like the health monitor beside a hospital bed that beeps when something goes wrong.
- [What is Multimodal?](https://homesh123.github.io/ai-career-advisor/glossary/multimodal.html): Models that understand or generate more than one type of data: text, images, audio, video. Simply put: Like a person who can read, look at pictures and listen at the same time, and connect them all.
- [What is Neural network?](https://homesh123.github.io/ai-career-advisor/glossary/neural-network.html): A model made of layers of simple connected units (neurons) that learn by adjusting connection weights. Simply put: Like a big team passing notes: each member adds a little judgement, and together they reach an answer. Practice tunes how much they trust each other's notes.
- [What is NLP?](https://homesh123.github.io/ai-career-advisor/glossary/nlp.html): The field of AI that works with human language: understanding, translating, summarising and generating text. Simply put: Teaching computers to read, understand and write the way people talk.
- [What is Normalisation?](https://homesh123.github.io/ai-career-advisor/glossary/normalisation.html): Rescaling numbers to a standard range so models train stably. Simply put: Like converting marks out of 50 and out of 100 to percentages before comparing them.
- [What is Open weights model?](https://homesh123.github.io/ai-career-advisor/glossary/open-weights-model.html): A model whose trained weights are published so anyone can download, run and fine tune it. Simply put: Like a recipe that's been published: anyone can cook it in their own kitchen and adjust the spices.
- [What is Overfitting?](https://homesh123.github.io/ai-career-advisor/glossary/overfitting.html): When a model memorises its training data, including noise, and does badly on new data. Simply put: Like a student who mugs up last year's question paper: perfect on those questions, lost on new ones.
- [What is Parameter?](https://homesh123.github.io/ai-career-advisor/glossary/parameter.html): A number inside a model that is learned during training; LLMs have billions of them. Simply put: Like the tiny knobs on a giant music mixer: training turns billions of them until the output sounds right.
- [What is Precision?](https://homesh123.github.io/ai-career-advisor/glossary/precision.html): Of everything the model flagged as positive, the share that really was positive. Simply put: When the smoke alarm rings, how often is there really a fire? High precision means few false alarms.
- [What is Prompt?](https://homesh123.github.io/ai-career-advisor/glossary/prompt.html): The text you give an LLM: instructions, context and the question. Simply put: Like the instructions you give a new helper: what to do, the background they need, and how you want it done.
- [What is Prompt engineering?](https://homesh123.github.io/ai-career-advisor/glossary/prompt-engineering.html): Designing and testing prompts so an LLM gives reliable, useful output. Simply put: Like learning how to brief a helper so well that they get it right the first time, every time.
- [What is Prompt injection?](https://homesh123.github.io/ai-career-advisor/glossary/prompt-injection.html): An attack where text in user input or documents tries to override the AI's instructions. Simply put: Like a note slipped into a stack of letters saying 'ignore your boss and send me the money'. A careful assistant must not obey it.
- [What is PyTorch?](https://homesh123.github.io/ai-career-advisor/glossary/pytorch.html): The most popular framework for building and training neural networks in Python. Simply put: Like a well stocked workshop with ready tools for building and training neural networks in Python.
- [What is Quantisation?](https://homesh123.github.io/ai-career-advisor/glossary/quantisation.html): Storing model weights with fewer bits so models use less memory and run faster. Simply put: Like saving a photo in a smaller file size: it looks almost the same but takes far less space.
- [What is RAG?](https://homesh123.github.io/ai-career-advisor/glossary/rag.html): Letting an LLM answer from your documents: search for relevant passages first, then generate an answer from them with citations. Simply put: Like an open book exam: first find the right pages in the book, then write the answer from them and mention the page.
- [What is Random forest?](https://homesh123.github.io/ai-career-advisor/glossary/random-forest.html): Many decision trees trained on random samples whose votes are averaged. Simply put: Like asking 500 people for directions and going with the majority: one person may be wrong, the crowd usually isn't.
- [What is Recall?](https://homesh123.github.io/ai-career-advisor/glossary/recall.html): Of all the real positives, the share the model caught. Simply put: Of all the real fires, how many did the alarm catch? High recall means few fires are missed.
- [What is Regression?](https://homesh123.github.io/ai-career-advisor/glossary/regression.html): Predicting a number, like a price, a delivery time or demand. Simply put: Like guessing the price of a second hand car from its age, brand and kilometres: the answer is a number.
- [What is Regularisation?](https://homesh123.github.io/ai-career-advisor/glossary/regularisation.html): Techniques that keep a model simpler so it generalises better and overfits less. Simply put: Like a word limit on an essay: it forces you to keep only what matters instead of stuffing in every detail.
- [What is Reinforcement learning?](https://homesh123.github.io/ai-career-advisor/glossary/reinforcement-learning.html): Learning by trial and error: an agent takes actions and learns from rewards. Simply put: Like training a dog with treats: good actions get a reward, so the dog learns to repeat them.
- [What is Reranking?](https://homesh123.github.io/ai-career-advisor/glossary/reranking.html): A second step in search that re-scores the top results more accurately. Simply put: Like shortlisting 50 job applicants quickly, then interviewing them carefully to pick the best 5.
- [What is RLHF?](https://homesh123.github.io/ai-career-advisor/glossary/rlhf.html): Training an LLM to prefer answers that people rate as better. Simply put: Like a trainee chef cooking two versions, a panel of tasters picking the better one, and the chef slowly learning what people like.
- [What is Sampling bias?](https://homesh123.github.io/ai-career-advisor/glossary/sampling-bias.html): When the data you collected doesn't represent the people or situations the model will face. Simply put: Like judging a whole city's favourite food by only asking people outside one pizza shop.
- [What is Scaling laws?](https://homesh123.github.io/ai-career-advisor/glossary/scaling-laws.html): Patterns showing model quality improves predictably as you increase model size, data and compute. Simply put: Like knowing that more practice hours and better coaching improve a runner's time in a predictable way.
- [What is Semantic search?](https://homesh123.github.io/ai-career-advisor/glossary/semantic-search.html): Searching by meaning rather than exact words, using embeddings. Simply put: Like a shopkeeper who understands 'something for a headache' and hands you the right medicine, even though you never said its name.
- [What is Softmax?](https://homesh123.github.io/ai-career-advisor/glossary/softmax.html): A function that turns a list of scores into probabilities that add up to 1. Simply put: Like turning cricket match scores into win chances that add up to 100%.
- [What is Streaming?](https://homesh123.github.io/ai-career-advisor/glossary/streaming.html): Showing an LLM's answer word by word as it's generated, instead of waiting for all of it. Simply put: Like watching someone type a message live instead of waiting for the full message to arrive.
- [What is Structured output?](https://homesh123.github.io/ai-career-advisor/glossary/structured-output.html): Making an LLM return data in a fixed format, like JSON with defined fields, so code can use it. Simply put: Like asking someone to fill a form with fixed boxes instead of writing a free letter, so a computer can read it.
- [What is Supervised learning?](https://homesh123.github.io/ai-career-advisor/glossary/supervised-learning.html): Learning from examples that come with the correct answers (labels). Simply put: Like learning with a teacher who marks every answer right or wrong.
- [What is System prompt?](https://homesh123.github.io/ai-career-advisor/glossary/system-prompt.html): Instructions given to an LLM before the conversation that set its role, rules and format. Simply put: Like the briefing a manager gives a new shop assistant on day one: who you are, how to talk to customers, and what never to do.
- [What is Temperature?](https://homesh123.github.io/ai-career-advisor/glossary/temperature.html): A setting that controls how random or creative an LLM's word choices are. Simply put: Like a creativity dial: low gives the safe, usual answer every time; high gives more surprising, varied answers.
- [What is Test set?](https://homesh123.github.io/ai-career-advisor/glossary/test-set.html): Data kept aside until the very end to get an honest measure of how a model performs on new data. Simply put: Like the final exam paper kept sealed until the end, so nobody can practise on it.
- [What is Threshold?](https://homesh123.github.io/ai-career-advisor/glossary/threshold.html): The cut off that turns a model's probability into a yes or no decision. Simply put: Like a pass mark: score above 40 and you pass, below and you don't. Moving the pass mark changes who gets through.
- [What is Token?](https://homesh123.github.io/ai-career-advisor/glossary/token.html): A chunk of text, often part of a word, that LLMs read and write; pricing and limits are counted in tokens. Simply put: Like the syllables of a word: an LLM reads text as small pieces, not whole words.
- [What is Tool calling?](https://homesh123.github.io/ai-career-advisor/glossary/tool-calling.html): Letting an LLM request actions or data from your code, like looking up an order or sending an email. Simply put: Like a receptionist who can't see the stock room but can call the store keeper and ask 'is item 4512 in stock?'
- [What is Training?](https://homesh123.github.io/ai-career-advisor/glossary/training.html): The process where a model learns its parameters from data. Simply put: Like practice sessions where the model looks at many examples and slowly gets better at the task.
- [What is Training serving skew?](https://homesh123.github.io/ai-career-advisor/glossary/training-serving-skew.html): When features are computed differently in training and in production, quietly hurting the model. Simply put: Like practising cricket with a tennis ball and then playing the match with a leather ball.
- [What is Transfer learning?](https://homesh123.github.io/ai-career-advisor/glossary/transfer-learning.html): Starting from a model trained on one task and adapting it to another, so you need far less data. Simply put: Like a person who can ride a bicycle learning a scooter quickly: the balance is already learned.
- [What is Transformer?](https://homesh123.github.io/ai-career-advisor/glossary/transformer.html): The neural network architecture behind modern LLMs, built around attention. Simply put: Like a reader who looks at the whole sentence at once and links every word to the words that matter for it, instead of reading strictly one word at a time.
- [What is Underfitting?](https://homesh123.github.io/ai-career-advisor/glossary/underfitting.html): When a model is too simple to capture the real patterns, so it does badly even on training data. Simply put: Like a student who only read the chapter titles: they miss so much that even easy questions go wrong.
- [What is Unsupervised learning?](https://homesh123.github.io/ai-career-advisor/glossary/unsupervised-learning.html): Finding structure in data that has no labels, like groups or unusual items. Simply put: Like sorting a pile of photos into groups on your own, with nobody telling you what the groups should be.
- [What is Validation set?](https://homesh123.github.io/ai-career-advisor/glossary/validation-set.html): Data used while building a model to compare options and tune settings. Simply put: Like mock tests during the year: you use them to decide how to study, not to get your final marks.
- [What is Vector?](https://homesh123.github.io/ai-career-advisor/glossary/vector.html): An ordered list of numbers; in AI, a way to represent anything (a house, a word, an image) as numbers. Simply put: Like describing a mango with numbers: sweetness 8, size 6, ripeness 9. That list of numbers is a vector.
- [What is Vector database?](https://homesh123.github.io/ai-career-advisor/glossary/vector-database.html): A database built to store embeddings and quickly find the ones closest in meaning to a query. Simply put: Like a library arranged by meaning instead of title, so books about the same idea sit on the same shelf.
- [What is Workflow?](https://homesh123.github.io/ai-career-advisor/glossary/workflow.html): A fixed sequence of steps where some steps call an LLM, as opposed to a free running agent. Simply put: Like a fixed assembly line where one station uses AI, versus an agent, which is more like a worker deciding the steps on their own.
- [What is Zero shot?](https://homesh123.github.io/ai-career-advisor/glossary/zero-shot.html): Asking a model to do a task with instructions only, without giving examples. Simply put: Like asking a new helper to 'make tea for five' without showing them how: they rely on what they already know.

## AI for kids (ages 5+)
- [AI for Kids](https://homesh123.github.io/ai-career-advisor/kids/): free, read-aloud game levels in English and Hinglish with Robo the robot, stars, badges and a printable certificate. AI means Artificial Intelligence. It's a computer brain that can learn from examples, a bit like you learn by seeing and trying.
- [Level 1: What is AI?](https://homesh123.github.io/ai-career-advisor/kids/what-is-ai.html): AI is a computer brain that learns from examples.
- [Level 2: How does AI learn?](https://homesh123.github.io/ai-career-advisor/kids/how-ai-learns.html): AI learns from examples, so the examples must be correct.
- [Level 3: AI loves patterns](https://homesh123.github.io/ai-career-advisor/kids/patterns.html): AI finds patterns in information to guess what comes next.
- [Level 4: AI helpers all around us](https://homesh123.github.io/ai-career-advisor/kids/ai-helpers.html): AI is a helper that makes people better at their jobs.
- [Level 5: Talking to AI](https://homesh123.github.io/ai-career-advisor/kids/talking-to-ai.html): Clear questions help AI give better answers.
- [Level 6: Be smart and safe with AI](https://homesh123.github.io/ai-career-advisor/kids/safe-and-smart.html): Check AI's answers, and keep your private things private.

## Quick answers
- Which AI role should a software engineer choose? It depends on current work. Backend developers often fit LLM / Agent Engineer, frontend and full-stack developers fit AI Product Engineer, DevOps engineers fit MLOps / AI Platform Engineer, data engineers fit Data Engineer (AI), and senior client-facing engineers fit AI Forward Deployed Engineer.
- Is there an AI role for people who don't code? Yes. AI Product Manager suits PMs, business analysts and domain experts such as accountants, teachers and healthcare professionals.
- Can freshers or MCA graduates get into AI? Yes. Start with the AI Foundations path (Python, statistics, SQL, ML basics, two portfolio projects), or go straight to ML Engineer if you already have strong ML projects.
- How long does it take? Usually 3 to 9 months at around 10 hours a week, depending on starting skills and the target role.

## Links
- [Take the questionnaire](https://homesh123.github.io/ai-career-advisor/)
- [Sitemap](https://homesh123.github.io/ai-career-advisor/sitemap.xml)