Advanced RAG Architecture & LLM Fine-Tuning
Engineering high-precision retrieval systems and domain-specific models for the enterprise.
Go beyond basic prompting. Learn to build enterprise-grade Retrieval-Augmented Generation (RAG) systems and fine-tune Large Language Models (LLMs) on private datasets with architectural precision.
- Level
- Advanced
- Status
- Available
- Duration
- 6 Modules
- Instructor
- Hoan Do (EPFL Master, ex-Paymentology Lead)
Curriculum
The RAG Fundamentals
Duration: 1h 45m
- Why RAG? The Context Problem — 25 min
- Chunking Strategies at Scale — 40 min
- Lab: Building Your First RAG Pipeline — 40 min
Vector Engineering
Duration: 2h
- Deep Dive: Vector Databases — 45 min
- Understanding Embeddings & Similarity — 45 min
- Optimizing for Low-Latency Retrieval — 30 min
Retrieval Optimization
Duration: 1h 45m
- Hybrid Search & Re-ranking — 35 min
- Query Transformation & Routing — 40 min
- Metadata Filtering — 30 min
RAG Evaluation
Duration: 1h 30m
- Measuring Faithfulness & Relevance — 45 min
- TruLens for Production Monitoring — 45 min
Fine-Tuning Deep-Dive
Duration: 2h 15m
- RAG vs. Fine-Tuning: When to Use Each — 30 min
- PEFT, LoRA, and QLoRA — 50 min
- Dataset Preparation for Instruct-Tuning — 55 min
Production Pipelines
Duration: 2h
- The Staff Engineer Perspective — 40 min
- CI/CD for Data & Model Weights — 50 min
- Final Project: Enterprise RAG + Fine-Tuned Model — 30 min