Forward Deployed

Agentic RAG: sections and modules

T9

Agentic RAG

Retrieval inside a control loop: grading and rewriting, CRAG and Self-RAG, adaptive routing, RAG as a tool, and feedback loops.

Requirement: Required
For whom: All technical paths
Notebooks: 16
Before this track: Advanced Agents
By: Sourav Banerjee

Objectives

Add grading, rewriting and fallback to a retrieval pipeline
Decide when self-correction is worth its extra calls

Retrieval that checks and corrects itself is how production RAG handles the questions a baseline gets wrong.

T9.1
Agentic RAG Foundations: Grading & Query Rewriting
5 lessons
5 lessons
T9.2
Self-Correcting RAG: CRAG & Self-RAG
5 lessons
5 lessons
T9.3
Adaptive RAG & Query Routing
6 lessons
6 lessons
T9.4
RAG as an Agent Tool
5 lessons
5 lessons
T9.5
Reliability, Feedback Loops & Diversity
5 lessons
5 lessons

Exercise: Add a relevance grader and a web fallback to the baseline, and measure what changes
Done when: You can answer every module's Interview questions without notes