A practical loop for building AI agents with Python
Start with a small, inspectable agent loop that chooses tools, observes results, and knows when to stop.
The knowledge library
Articles
Practical guides, tutorials, experiments and things I've learned while building.
8 articles
RAG Architecture Deep Dive: Vector Embeddings, Hybrid Search & Reranking
A production engineering guide to building accurate Retrieval-Augmented Generation pipelines using dense embeddings, BM25 keyword matching, and cross-encoder rerankers.
RAG, explained as a retrieval pipeline
A grounded introduction to chunking, embeddings, retrieval, and the evaluation work that makes RAG useful.
Model Context Protocol (MCP) Guide: Standardizing AI Tool Integration
Everything developers need to know about Anthropic's Model Context Protocol (MCP) to connect LLMs to data sources, local systems, and external APIs.
MCP for developers: connect tools through a shared protocol
Understand the host, client, and server roles in Model Context Protocol and how to think about tool boundaries.
Modern Async Python for AI Engineers: Handling High Concurrency
Master asyncio, connection pooling, and token streaming to build high-throughput AI backends capable of serving thousands of simultaneous requests.
Evaluating LLM Applications: Metrics, Evals & Automated CI Pipelines
How to move past the 'vibe check' by creating robust evaluation datasets, deterministic unit checks, and LLM-as-a-judge pipelines.
System Design for AI Systems: Caching, Rate Limits & State Persistence
Architectural blueprints for scaling agentic systems to production: semantic caching, backpressure handling, distributed state, and failure recovery.