Take-Home Assignment Survival: Scope, Automated Tests, and README Craft (2026)
9/12/202615 min read
11 articles tagged with Testing
The Depends() system is the reason FastAPI test suites for LLM services can run in milliseconds without ever touching a real model API. Here's how to structure it so that stays true.
Unit tests don't work on non-deterministic outputs. Here's how to build an evaluation pipeline that catches regressions in an LLM app the same way a test suite catches them in ordinary code.
You can't put a headset in a CI runner. Here's the actual stack I use to test WebXR code — emulated sessions for local development, mocked navigator.xr for unit tests, and where real devices remain non-negotiable.