Course Objective & Scope
AI coding tools that shine on greenfield demos often fail in hundred-thousand-line enterprise monorepos—putting code in the wrong architectural layer, reinventing existing utilities or forgetting constraints as the context window fills up. This course covers repository-scale context engineering so coding agents navigate and modify large codebases accurately.
Hierarchical memory files (AGENTS.md, CLAUDE.md, .cursor/rules), AST (tree-sitter) and LSP symbol indexing, hybrid lexical + semantic code search, explicit module boundaries, MCP context providers and context compaction/handoff patterns are practised on a multi-package repository.
Who Should Attend?
- Senior and staff software engineers and software architects
- Platform, DevEx and developer-tooling engineers
- Tech leads managing large or monorepo codebases
- Prerequisite: hands-on experience with large codebases and AI coding tools
Key Takeaways
- Structure repositories so agents locate the right modules and patterns with minimal token spend
- Design hierarchical, path-scoped instruction files from repo root down to individual packages
- Combine AST chunking, symbol graphs and hybrid search for high-precision code retrieval
- Prevent context poisoning and context bloat through structured compaction and subagent isolation
- Expose live database schemas, design tokens and internal SDK docs via MCP context servers
Curriculum
The Context Challenge in Large Repos and Context Rot
- Why giant context windows are not enough: lost-in-the-middle effects and attention dilution
- Token budget economics across static instructions, dynamic retrieval and tool outputs
Agent-Navigable Repository Structure and Naming
- Predictable directory layouts, explicit public module APIs and expressive type signatures
- Making shared utilities and canonical patterns easy for agents to discover
Hierarchical Memory: AGENTS.md, CLAUDE.md and Scoped Rules
- Progressive context disclosure across root, service and package-level files
- Path-scoped rule activation and pointing agents to canonical golden examples
Code Indexing: ASTs, LSPs and Hybrid Retrieval
- Tree-sitter AST chunking and generating compact repository symbol maps
- Combining lexical search (ripgrep), LSP call/reference graphs and vector embeddings
Dynamic MCP Context Providers and Session Compaction
- Serving internal library docs, DB schemas and API contracts through MCP resources
- Isolating exploration from implementation and handing off state via structured artifacts
Other Courses in This Category
-
Detecting OWASP Top 10 flaws, broken access control, hallucinated packages (slopsquatting), leaked secrets and insecure defaults in AI- and human-written code using SAST (Semgrep, CodeQL) and LLM-assisted security review.
- Live Course
- 5 modules
-
Governance training for engineering organisations covering zero-data-retention (ZDR), IP and open-source licence compliance, approved model/MCP registries, risk tiering and developer policies for AI-assisted coding.
- Live Course
- 5 modules
-
End-to-end training on redesigning the Software Development Life Cycle for the AI era—resolving review, QA and security bottlenecks across requirements, coding, PR review, CI/CD and operations, with DORA/SPACE metrics.
- Live Course
- 5 modules
-
Extracting business rules from COBOL, PL/SQL stored procedures and legacy Java/.NET monoliths, applying the Strangler Fig pattern, verifying parity with golden-master/dual-run tests and migrating safely to modern architectures.
- Live Course
- 5 modules
-
Training for QA and test engineers on testing non-deterministic LLM, RAG and agent applications using golden datasets, RAG evaluation metrics, LLM-as-a-Judge, prompt-injection red teaming and CI regression gates.
- Live Course
- 5 modules
-
Training for SRE, DevOps and IT operations teams on alert noise reduction, log/metric/OpenTelemetry trace correlation, recent-change analysis, human-approved runbook execution and post-mortem automation.
- Live Course
- 5 modules
Get in touch
