When Q4 Passes Latency but Fails Quality
See why a Q4 model that fits and meets latency can still fail a release gate—and why incomplete evidence must not select a quantization.
Read noteThe complete notebook
All articles
Deep dives and field notes on intelligent systems, search, and backend platforms.
Serve a 3B LLM on CPU With llama.cpp
Understand CPU inference from first principles: autoregressive decode, memory bandwidth, KV state, context allocation, scheduling, and admission control.
Read noteBenchmark LLM Serving Without Lying to Yourself
Understand inference performance as a workload-dependent surface, then test saturation and goodput with a controlled local benchmark.
Read noteBuild a Fully Offline RAG App in Your Browser
Build a browser-only RAG application that locally parses TXT, Markdown, PDF, and DOCX files and answers questions after the network is disabled.
Read noteThe State of Agentic AI in 2026: What Changed, What Matters, and How to Start
A practical field guide to the capabilities, protocols, risks, and minimum path for building a first useful AI agent in 2026.
Read note