The Library
Insights, strategies, and deep dives into the world of AI building and geeky living.
EvaluationSearchAI
A Benchmark Can Pass Every Validity Check and Still Fail You
Why a benchmark can pass every validity check you throw at it, and still be statistically meaningless in practice.
Aug 8, 20268 min read
Read More
EvaluationSearchAI
My retrieval benchmark passed the replace-all-vectors-with-noise test.
How to avoid certifying a benchmark you never tested. A deep dive into retrieval evals, noise floors, and why one check is never enough.
Aug 5, 20266 min read
Read More
AIArchitectureMemory
Navigating AI Memory & Context Window Issues
Deep dive into the challenges of managing context windows in production AI applications and how to overcome them.
Feb 8, 20268 min read
Read More
DevelopmentAI FrameworksStrategy
Frameworks & Strategies to Build Scaleable AI Apps
A comprehensive guide to the best frameworks and architectural strategies for building agentic AI systems today.
Feb 7, 202610 min read
Read More