Research28 Aug 2026 · 7 min
The tail is the product
Why we throw away aggregate F1 on tagging benchmarks, and what we report instead. A short argument with a long table attached.
Blog
Method write-ups, evaluation arguments and the occasional account of six weeks spent on something that did not work.
5 posts
Research19 Aug 2026 · 9 min
Regions, not repairs
Inpainting treats a mask as damage. PrismStream treats it as composition. Notes from six weeks of getting that distinction to survive training.
Data6 Aug 2026 · 6 min
Filters are opinions, so ship the score
On building Aurora: why every quality filter in the pipeline emits a column instead of deleting a row, and what that costs in storage.
Infrastructure21 Jul 2026 · 5 min
Scheduling for four GPUs
Kubernetes answers a question small labs are not asking. Notes on writing a scheduler whose main feature is explaining itself.
Lab4 Jul 2026 · 2 min
Hello world
The first post. Mostly here so the blog has something to render while the real writing gets started.