Master’s Program
GenAI Product Builder
RAG, agents and evaluation — build GenAI that actually works
Overview
Learn how to build GenAI products with measurable quality. You’ll implement retrieval-augmented generation, prompt patterns, evaluation loops, and guardrails — and ship a working demo with clear metrics.
Key outcomes
- Build RAG systems with evaluation-driven iteration
- Implement safety guardrails and quality checks
- Create stakeholder-friendly demos and documentation
Format
- Project-first approach
- Remote-friendly (worldwide cohorts)
Tools
- RAG concepts
- Vector DB concepts
- APIs
- Evaluation strategy
Detailed program
Build GenAI products with measurable quality: RAG, evaluation loops, reliability patterns and guardrails — shipped as a working demo.
Module 1 — GenAI Fundamentals & Product Constraints
Understand what LLMs can and cannot do in real products.
1–2 weeks
Module 1 — GenAI Fundamentals & Product Constraints
Understand what LLMs can and cannot do in real products.
What you’ll learn
- LLM capabilities vs limits: hallucinations, context windows, latency
- Prompting patterns: instructions, role, examples, constraints
- Structured outputs (schemas), parsing, and failure handling
- Product constraints: cost, privacy, data access, UX expectations
Skills you’ll gain
Module 2 — RAG Architecture (Chat with your data)
Implement Retrieval-Augmented Generation with strong retrieval quality.
2–3 weeks
Module 2 — RAG Architecture (Chat with your data)
Implement Retrieval-Augmented Generation with strong retrieval quality.
What you’ll learn
- Embeddings and semantic search intuition
- Chunking strategy, overlap, metadata, and indexing
- Retrieval strategies: top-k, filtering, re-ranking basics
- Grounded answers with citations and source linking
- Freshness and update strategy for knowledge bases
Skills you’ll gain
Module 3 — Evaluation & Quality Measurement
Measure quality and avoid shipping ‘vibes-based’ GenAI.
2 weeks
Module 3 — Evaluation & Quality Measurement
Measure quality and avoid shipping ‘vibes-based’ GenAI.
What you’ll learn
- Define success metrics: accuracy, groundedness, usefulness, safety
- Offline evaluation: test sets, golden answers, regression tests
- Human evaluation basics: rubrics, review workflows
- Failure analysis: hallucinations, retrieval misses, prompt brittleness
- Iterate with measurable improvements
Skills you’ll gain
Module 4 — Guardrails, Safety & Reliability
Add protection layers that make GenAI usable in production contexts.
1–2 weeks
Module 4 — Guardrails, Safety & Reliability
Add protection layers that make GenAI usable in production contexts.
What you’ll learn
- Safety risks overview: prompt injection, data leakage, toxic content
- Guardrails patterns: input filters, output constraints, policy checks
- Prompt injection awareness + mitigation patterns
- Fallback strategies: refusal, safe completion, escalation to human
Skills you’ll gain
Module 5 — GenAI Product Delivery (API + Demo + Docs)
Ship a complete product: API + UI demo + documentation.
2–3 weeks
Module 5 — GenAI Product Delivery (API + Demo + Docs)
Ship a complete product: API + UI demo + documentation.
What you’ll learn
- Build a simple API for inference (request/response discipline)
- Basic performance mindset: latency, caching, batching awareness
- UI demo: clear user journeys, examples, and limitations section
- Documentation: how it works, known limitations, evaluation results
Skills you’ll gain
Capstone — RAG Product (Portfolio-grade)
Deliver a working RAG app with evaluation results and guardrails.
2–3 weeks
Capstone — RAG Product (Portfolio-grade)
Deliver a working RAG app with evaluation results and guardrails.
What you’ll learn
- Define a use case (support assistant, knowledge bot, internal search)
- Implement retrieval + citations + evaluation loop
- Add guardrails + failure handling + metrics summary (simple)
- Finalize demo + documentation + portfolio presentation
Skills you’ll gain