AI Engineering · Live cohort

AI Engineering Programme

Build, evaluate and ship LLM applications (RAG, agents, fine-tuning, LLMOps) and get ready for AI engineering interviews.

Live · Sat + Sun, 4–6 PM IST
12 weeks
Entry to mid level (0–6 yrs) · Python required

Apply for this cohort Try a free bootcamp first

What's included

What you get in every cohort.

Live sessions

12 weeks of live classes plus weekly doubt-clearing. All recorded.

10 mock interviews

Realistic rounds with written feedback from a mentor.

500+ interview questions

A curated bank, organised by topic and difficulty.

1:1 mentoring

Scheduled time with a mentor for code, career and resume reviews.

Certificate

A verifiable certificate on completing the programme and capstone.

About the programme

What you'll be able to do

A 12-week live, mentor-led cohort for software and data engineers moving into AI engineering. Every weekend has two 2-hour live sessions: Saturday to learn the concept, Sunday to build it. You'll ship a retrieval-augmented (RAG) app, an agent with tools, an evaluation harness and a fine-tuned open model, then bring them together in a production-ready capstone. Mock interviews start in week 2 and build to 10, backed by a 500-question bank (LLM fundamentals, RAG and agents, AI system design and coding) and 1:1 reviews with a mentor.

Build LLM features in Python with prompting, structured outputs and tool calling
Design RAG systems that retrieve the right context, cite it, and hold up under real evaluation
Build agents and multi-step workflows that fail safely
Choose between prompting, RAG and fine-tuning, and fine-tune an open model with LoRA
Run LLM features in production: latency, cost, observability and guardrails
Go into AI engineering interviews having done 10 mock interviews with written feedback

Weekly rhythm

A typical week

SaturdayLive class, 4–6 PM IST (2 hrs)
SundayLive build session, 4–6 PM IST (2 hrs)
WeekdaysLab + assignment (~3 hrs, at your own pace)
Most weeksMock interview + written feedback (10 in total)

Syllabus

Week by week

Every week pairs live teaching with something you build or practise.

Week 1LLM foundations for engineers

Sat: how LLMs work: tokens, context windows, sampling, cost
Sun: calling model APIs from Python: streaming, retries, rate limits
Choosing a model: quality, latency and price trade-offs

This week: CLI assistant with streaming and per-call cost logging

Week 2Prompt engineering & structured outputs

Sat: prompt patterns, few-shot examples, system prompts
Sun: JSON and schema-validated outputs, function calling
Versioning and testing prompts like code

This week: Information-extraction service with schema validation + mock interview #1: LLM fundamentals

Week 3Embeddings & vector search

Sat: embeddings, similarity and chunking strategies
Sun: vector stores (pgvector, FAISS) and hybrid search with BM25
Metadata filtering and index design

This week: Semantic search over a documentation set + mock interview #2: Python coding

Week 4Retrieval-augmented generation (RAG)

Sat: the RAG pipeline end to end, context packing, citations
Sun: reranking, query rewriting and handling "I don't know"
Common RAG failure modes and how to spot them

This week: RAG Q&A app over real documents, with citations

Week 5Evaluating LLM systems

Sat: building eval datasets, retrieval metrics (recall@k, MRR)
Sun: LLM-as-judge, regression tests and evals in CI
Measuring hallucination and faithfulness

This week: Evaluation harness for your week-4 RAG app + mock interview #3: RAG design

Week 6Tool use & agents

Sat: tool calling, agent loops, planning and memory
Sun: building an agent with tools and the Model Context Protocol (MCP)
Failure modes: loops, bad tool calls, runaway cost

This week: Agent that completes a multi-tool task safely + mock interview #4: coding

Week 7Multi-step workflows

Sat: workflows vs agents, state machines, human-in-the-loop
Sun: orchestration frameworks vs plain code, retries and checkpoints
Multi-agent patterns and when not to use them

This week: Production-style workflow with approvals + mock interview #5: AI system design

Week 8Fine-tuning & open models

Sat: prompting vs RAG vs fine-tuning, and dataset preparation
Sun: LoRA/QLoRA fine-tuning of a small open model
Serving open models: vLLM, Ollama, quantisation

This week: LoRA fine-tune evaluated against the base model + mock interview #6: ML fundamentals

Week 9LLMOps in production

Sat: latency and cost: caching, batching, model routing
Sun: tracing and observability, deploying with FastAPI and containers
Guardrails: prompt injection, PII and output validation

This week: Deployed RAG service with tracing and guardrails + mock interview #7: AI system design

Week 10AI system design

Sat: designing a support copilot and an enterprise search assistant
Sun: trade-offs at scale: freshness, cost, privacy, evaluation
Whiteboarding an AI system the way interviews test it

This week: Design doc reviewed by a mentor + mock interview #8: system design

Week 11Interview sprint

Sat: LLM concepts and RAG questions under time pressure
Sun: coding + AI system design drills from the question bank
Behavioural stories that land in AI engineering loops

This week: Mock interview #9: full loop

Week 12Capstone & final mocks

Sat: capstone demos with mentor feedback
Sun: resume, LinkedIn and GitHub portfolio polish
Final full-loop mock interviews

This week: Capstone demo + mock interview #10 · certificate on completion

Your mentor

Taught by an engineer who does this work today

Abhi Sharma

Data Engineer, Google

Abhi is a Data Engineer at Google, working on the same class of cloud data infrastructure these programmes teach. He is Google Cloud certified as a Professional Data Engineer and a Professional Machine Learning Engineer, and writes on data engineering, ML and system design for 4,000+ engineers.

GCP Professional Data EngineerGCP Professional ML Engineer

Mentors on: Data pipelines · Airflow & Kafka · Data warehousing · ML engineering · System design

LinkedIn profile

Apply

Apply free. Pay only if you join.

Apply — two minutes, free.
Talk to a mentor — we check the programme fits your goals and answer your questions.
Confirm your seat — pay using the options we share on the call, then join the cohort.

We don't promise jobs or salary hikes. We promise live teaching, honest feedback and real practice.

Questions

Before you apply

How much time does it take each week?

Plan for about 6–8 hours: 3–4 hours of live sessions, depending on the programme, plus labs and practice. Every programme is designed around a full-time job.

What if I miss a live session?

Every live class is recorded and available in your dashboard, and you can bring questions to the next live session.

Who are the mentors?

Engineers currently working at top product companies. You'll see each mentor's profile on the programme page before you apply.

How does payment work?

Apply first. It's free and takes two minutes. We'll get in touch to confirm fit and share the payment options before you pay anything.

Do you guarantee a job or a salary hike?

No, and be wary of anyone who does. We give you the skills, the practice and honest feedback from engineers at top product companies. The outcome depends on the work you put in.