Writing
Latest insights
Practical guides and insights on implementing AI systems that deliver real value. No hype, no fluff.
The Model Is a Commodity Now: A Year of Building Around It
Notes from June 2025 to June 2026. The best model in the world got pulled offline in under a week, which made something click: the model is rented, the system around it is what you own. What I built to convince myself, in the open.
Your LLM-as-Judge is Lying to You
The uncomfortable truth about automated LLM evaluation systems: position bias, verbosity worship, and the models that prefer their own outputs
Agent Design Patterns and Architectures for LLM-Based AI Agents (2025)
A comprehensive taxonomy of design patterns for modern LLM-based AI agents, from single-agent loops to complex multi-agent systems
Introduction to LLM Evaluations
Understanding how to effectively evaluate Large Language Models for production AI systems
Building Effective RAG Systems: A Comprehensive Guide
Learn how to create production-ready Retrieval Augmented Generation (RAG) systems that scale effectively and deliver accurate results.
Evaluation Driven Development
A systematic approach to developing and improving LLM applications