AI-Assisted Coding | Session 2: Repository Context — RAG and MCP

AAI-Assisted Coding | Session 2: Repository Context — RAG and MCP

We offer a 50% discount to students; please complete this form to avail the discount.

How do you give an AI assistant the right repository context, and only the tools it should be allowed to use?

This session explores Retrieval-Augmented Generation (RAG) and the Model Context Protocol (MCP), covering embeddings, vector search, code and documentation chunking, metadata, access control and incremental indexing.

You will learn to improve retrieval using filtering, top-k selection, reranking, context budgets and citations, and evaluate retrieval quality using golden questions while considering latency and cost.

The session also introduces MCP hosts, clients and servers, the distinction between tools, resources and prompts, and the official MCP Python SDK. Security topics include least privilege, approvals, prompt-injection defence and audit logs.

This session is the second of a three-session professional training programme:

• Session 1 — Code Generation: Safe Prompts and Verified Change

• Session 2 — Repository Context: RAG and MCP

• Session 3 — Agentic AI: Agents, Workflows and Capstone

REGISTRATION FOR ALL 3 SESSIONS IS 150$ ONLY.

What You Will Learn

• Embeddings and vector search

• Chunking code and documentation with metadata

• Access control and incremental indexing

• Filtering, top-k selection and reranking

• Context budgets and citations

• Retrieval evaluation using golden questions

• Retrieval latency and cost

• MCP hosts, clients and servers

• When a capability should be a tool, a resource or a prompt

• The official MCP Python SDK

• Least-privilege access and approval controls

• Prompt-injection defence and audit logs

Why This Matters

AI assistants need relevant repository information to produce useful results, but access to context and tools must remain controlled.

This session develops a practical approach to repository retrieval and MCP integration, combining retrieval evaluation with explicit permission boundaries and security controls.

Ideal For

• Software engineers

• Technical leads

• QA engineers

• Software architects

• Engineers building repository-aware AI assistants

Basic programming ability, preferably in Python, and basic Git and command-line experience are required.

Training Format

• 4-hour live online training session

• Concise teaching blocks

• Four live demonstrations:

  • Demo 5: Parse a repository, embed it and store metadata-rich chunks in Qdrant.

  • Demo 6: Tune filters and check retrieved chunks and citations against golden questions.

  • Demo 7: Build an MCP server exposing one repository resource and one typed test-running tool.

  • Demo 8: Inspect the server in MCP Inspector and add path restrictions and approval controls before any side effect.

• A ten-minute break and five-minute quiz exactly at the two-hour mark

• Practical exercises using the sample repository

Tools and Technologies

Python, Qdrant, the official MCP Python SDK and MCP Inspector.

This session is Session 2 of the three-session AI-Assisted Coding and Code Generation Programme.