RAG Workshop
From Zero to Production
One-day intensive workshop. Build production-ready RAG pipelines hands-on, then master 9 advanced RAG patterns — Hybrid, Graph, Multi-Hop, Agentic, Self-Corrective, Multimodal & more.
17 exercises · 19 notebooks · Lifetime recording access
What You'll Learn
From your first hallucination demo to production-ready architecture — in one intensive day.
Why RAG Matters
See LLM hallucination firsthand, then fix it with RAG. Understand why RAG beats fine-tuning.
Chunking & Embeddings
Master fixed, recursive, and semantic chunking. Build vector indexes with FAISS and ChromaDB.
Build a RAG Pipeline
Load → Chunk → Embed → Store → Retrieve → Generate. End-to-end, from scratch.
Advanced Retrieval
HyDE, Multi-Query, BM25+Dense Hybrid Search, RRF Fusion, MMR diversity, Cross-Encoder Reranking.
9 RAG Patterns
Hybrid, Graph, Multi-Hop, Agentic, Self-Corrective, Multimodal, Hierarchical RAG — all covered.
Evaluation & Production
RAGAS metrics, semantic caching, PII scrubbing, guardrails, and the 6-layer production architecture.
Your Learning Journey
One intense day. Two distinct phases. You leave with working code and deep understanding.
Break the LLM
Exercise 1See GPT hallucinate on your own documents. Understand why RAG exists and when to use it vs. fine-tuning.
Chunking Strategies
Exercise 2 & 2bFixed-size, recursive, and semantic chunking. Learn to split documents so retrieval actually works.
Embeddings & Vector Stores
Exercise 3Turn text into vectors that capture meaning. Build indexes with FAISS, ChromaDB, and Pinecone.
Your First RAG Pipeline
Exercise 4Load → Chunk → Embed → Store → Retrieve → Generate. Build a complete Q&A bot end-to-end.
Advanced Retrieval
Exercise 5HyDE (hypothetical documents), Multi-Query expansion, and Cross-Encoder reranking to boost precision.
Hybrid Search + MMR
Exercise 5bCombine BM25 keyword search with dense embeddings using Reciprocal Rank Fusion. Maximize diversity with MMR.
RAGAS Evaluation
Exercise 6Measure your pipeline: Faithfulness, Relevancy, Context Precision & Recall. Find and fix the weakest link.
LangChain & LlamaIndex
FrameworksThe two dominant RAG frameworks compared. Build the same pipeline in both, understand when to pick which.
Hybrid RAG
Advanced BM25 + Dense fusion with score normalization and learned weights
Graph RAG
Knowledge graphs, entity extraction, relationship triples with NetworkX
Multi-Hop RAG
Bridge-fact questions, iterative retrieval, and query decomposition chains
Agentic RAG
ReAct pattern — LLM decides when to retrieve, which tools to use, and when to stop
Self-Corrective RAG
CRAG: grade retrieved docs, refine queries, and fallback to web search automatically
Multimodal RAG
Caption-first pattern — index images alongside text using vision models
Hierarchical RAG
Parent-child chunks, sentence windows, and 3-level summary pyramids
Evaluation Deep Dive
Golden datasets, quality-latency curves, cost vs. accuracy trade-offs
Production RAG
6-layer architecture: guardrails, caching, PII scrubbing, observability, and deployment
Test your knowledge with a live quiz, get your certificate, and take home all notebooks + resources.
Curriculum Deep Dive
8 hands-on exercises you code along, plus 9 advanced patterns taught in depth.
Break the LLM
- •Why LLMs hallucinate — knowledge cutoff, no access to private data
- •RAG vs. Fine-Tuning vs. Prompt Engineering — when to use each
- •Your first RAG motivation: see the problem, then fix it
Ask GPT about private documents — watch it hallucinate, then fix it with RAG.
Chunking Strategies
- •Fixed-size vs. Recursive Character Text Splitter (industry default)
- •Chunk size & overlap tuning — small (precise) vs. large (context-rich)
- •Semantic chunking: embedding-based splitting at topic boundaries
Compare 3 chunking strategies on the same document and measure retrieval quality.
Embeddings & Vector Stores
- •What embeddings are — vectors that capture meaning, not just keywords
- •Cosine similarity, embedding models (text-embedding-3-small)
- •FAISS (in-memory), ChromaDB (persistent), Pinecone/Weaviate (cloud-scale)
Build a vector index, query it, and see how 'refund' matches 'money back'.
Basic RAG Pipeline
- •The full pipeline: Load → Chunk → Embed → Store → Retrieve → Generate
- •Prompt engineering for grounded answers ('say I don’t know')
- •Top-k tuning, source citations, conversation memory
Build a complete Q&A bot and benchmark it against 10 test questions.
Advanced Retrieval
- •HyDE: generate hypothetical answer, embed that for better retrieval
- •Multi-Query: generate query variants to catch different phrasings
- •Cross-Encoder Reranking: retrieve 20, rerank to top 5
Implement HyDE + reranking and measure precision improvement.
Hybrid Search + MMR
- •BM25 (keyword) + Dense (semantic) — best of both worlds
- •Reciprocal Rank Fusion (RRF): combining rankings from multiple retrievers
- •MMR: balance relevance and diversity to avoid duplicate passages
Build a hybrid retriever with tunable alpha and measure recall improvement.
RAGAS Evaluation
- •4 metrics: Faithfulness, Answer Relevancy, Context Precision, Context Recall
- •LLM-as-judge: automated evaluation at scale without human annotation
- •Diagnostic loop: find lowest metric → fix specific cause → re-evaluate
Run RAGAS on your pipeline, get a baseline, improve one thing, measure the delta.
Hybrid RAG
Advanced BM25 + Dense fusion with score normalization and learned weights
Graph RAG
Knowledge graphs, entity extraction, relationship triples with NetworkX
Multi-Hop RAG
Bridge-fact questions, iterative retrieval, query decomposition
Agentic RAG
ReAct pattern — LLM decides when to retrieve and which tools to use
Self-Corrective RAG
CRAG: grade retrieved docs, refine queries, fallback to web search
Multimodal RAG
Caption-first pattern — index images alongside text with vision models
Hierarchical RAG
Parent-child chunks, sentence windows, 3-level summary pyramid
Evaluation Comparison
Golden datasets, quality-latency curves, cost vs. accuracy trade-offs
Production Readiness
6-layer architecture: guardrails, caching, PII scrubbing, observability
Your Workshop Toolkit
Three custom-built apps to accelerate your learning.
RAG Playground
ragplayground.futureproofindia.com
Interactive demos of all RAG patterns + LangChain & LlamaIndex cheatsheets
RAG Buddy
ragbuddy.futureproofindia.com
AI chatbot trained on the workshop curriculum — ask it anything during exercises
Live Quiz
ragplayground.futureproofindia.com/quiz
Fun interactive quiz covering all 17 exercises — test your knowledge with confetti
Your Instructor
Sridhar Jammalamadaka
Founder, Future Proof India | AI Architect
- •16 years in IT with deep expertise in enterprise AI, cloud architecture, and scalable systems
- •AI Architect — former Software Architect at Oracle, M.Tech from IIIT Bangalore
- •Trained 1,400+ professionals in Agentic AI, RAG, and LLM Engineering
- •Built 30+ AI agents and 50+ AI chatbots for enterprise clients across healthcare, finance, and legal domains
- •Trainer at IIIT Bangalore Executive Education & AI Consultant serving India, Dubai, and USA
- •Weekly Agentathon organizer — participants build and ship an AI agent every Saturday
Who Is This For?
Software Engineers
Want to add RAG capabilities to your applications and understand the full pipeline.
Tech Leaders & Architects
Need to evaluate RAG solutions, make architecture decisions, and lead AI initiatives.
Data Scientists & ML Engineers
Looking to move beyond model training into building retrieval-augmented AI systems.
AI Enthusiasts
Have basic Python skills and want to learn the most in-demand AI engineering skill of 2026.
What's Included
Prerequisites
Invest in Your AI Career
One day that transforms your ability to build production AI systems.
- ✓Full-day live workshop (9 hours)
- ✓17 exercises + 19 Colab notebooks
- ✓9 advanced RAG patterns taught
- ✓3 custom workshop apps (Playground, Buddy, Quiz)
- ✓Workshop recording (lifetime access)
- ✓Certificate of completion
- ✓1-week post-workshop support
- ✓Private Telegram Group access
GST included. Full refund available up to 24 hours before the event.
Frequently Asked Questions
Is this workshop online or offline?+
Will I get a recording?+
What if I'm a complete beginner in AI?+
What tools do I need?+
Can I get a refund?+
Will I be able to build a RAG app after this?+
What are the 9 advanced RAG patterns?+
Ready to Master RAG?
28 March 2026 · 10 AM – 7 PM IST · Live Online
Limited seats to ensure hands-on attention for every participant.
Register Now — ₹10,000Questions? Reach out to us:
📧 connect@futureproofindia.com · 📱 WhatsApp: +91 63603 19758