Philip Cheung · AI Engineering Lab
Published AI engineering posts as build-in-public evidence.
A compact view of Philip's published AI engineering posts, mapped to inference and serving, agents and harnesses, retrieval and data, evaluation loops, and AI product delivery.
Latest post
2026-08-11
Harness Hygiene: Pruning AI Context Bloat
Media coverage
31/31
Video-backed posts ready for direct inspection
Primary lens
Agents / Product / LLM Ops
Topic filters map posts to AI engineering evidence
Build map
Published posts
AI engineering posts organised across inference, agents, retrieval, evals and product delivery
31
posts
31
videos
5
lanes
Video album
Enterprise AI Agent Observability & Evaluation
build radar
Media-backed published posts with full article detail.
Evaluation discipline
Posts are sorted into eval and review lanes, showing how shipped AI work is tested before it is trusted.
Latency engineering
Inference and serving posts track the cost, latency and throughput decisions behind production LLM features.
Product judgement
Agent, retrieval and product posts show how AI capability turns into shipped, governed product surfaces without leaving philip.pm.