The blog
Thoughts on technology, things I'm learning, and whatever else catches my attention.
Featured post
· 5 min read
How AI coding agents changed my work: less implementation, more planning, review, and thinking about failure. I miss writing code, but I'm not going back.
AI Coding · Software Engineering · Personal
Build an eval from scratch: choose a task, put together a small dataset, write your scorers, and use the results to improve the next version.
AI Evals · Walkthrough · Python
What evals measure, how they differ from benchmarks and unit tests, and a worked example you can try yourself.
AI Evals · Fundamentals
A practical look at datasets, scorers, human review, and turning production failures into tests you can run again.
AI Evals · LLM Engineering · Observability · Quality
How we turn publisher PDFs into classroom questions at Wayground, using OCR, LangGraph, symbolic verification, and quality checks.
AI Platform · LangGraph · EdTech · Multimodal