AI in production
Shipping GenAI Features Without Breaking Core Systems
Integration patterns for reliable APIs, observability, and graceful fallback in existing systems.
Read articleNotes from engineering
Thinking through
the system.
Practical writing on GPU workloads, data infrastructure, and bringing new capabilities into existing software.
Data infrastructure
Scaling Vector Search Beyond a Billion Points
Design principles for partitioning, storage tiers, and recall-preserving performance at scale.
Read articleGPU computing
How GPU-First Architecture Cuts Model Latency
A practical guide to reducing inference latency with memory-aware design, batching, and end-to-end profiling.
Read article