Computational experiments and systems research

On the persistence of statistical anomalies in production machine learning pipelines and the limits of automated remediation.

Recurrent themes in software architecture

We examine the trade-offs between strict consistency models and high-availability distributed stores. The inquiry centers on failure domains that manifest only under sustained load.

Subsequent notes explore team topology in early-stage engineering organizations, arguing for small, autonomous squads anchored by first-principles technical literacy.

Receive quarterly dispatches

Delivered four times a year. No algorithms, no sponsors, direct to your inbox.