FOR ENGINEERS

Research.

Working notes and longer investigations — the unpolished thinking behind the shipped systems.

GPU · RUNTIMEWorking note · 2026

Scheduling LLM inference across consumer GPUs

The bottleneck isn't compute. It's what you choose to run, and when — a scheduler that picks the right backend per operation beats brute-force parallelism on hardware that was never meant to share a workload.

READ THE NOTE →
AGENTS · SYSTEMSDraft · 2026

Protocols for handing work between agents under partial context

READ THE NOTE →
OPERATING SYSTEMSNotes · 2026

Treating models as processes in an AI-native operating system

If a model can be scheduled, it can be a citizen of the kernel, not a guest of it — the OS primitives that manage processes (priority, preemption, resource limits) apply almost directly to managing models.

READ THE NOTE →
OBSERVABILITY · MLOPSWorking note · 2025

Statistical anomaly detection beats LLM-judged evals for production monitoring

READ THE NOTE →
SYSTEMS · CMSNotes · 2026

A block-based CMS is the right shape even when the only editor is you

Four client CMSes and one portfolio later: uniform blocks with dumb fields beat clever schemas every time someone (including future-you) has to actually edit content.

READ THE NOTE →
AGENTS · VOICEWorking note · 2026

Tool-calling over a voice WebSocket: keeping one contract for text and speech

READ THE NOTE →
AGENTS · RETRIEVALDraft · 2026

Deep repo knowledge for an on-site agent — without a vector database

For two dozen repos, pre-written docs beat embeddings: index once into a handful of LLM-authored documents (architecture, stack, per-module dives), then let the agent read titles and fetch whole docs on demand.

READ THE NOTE →