Decoding one token reads every weight out of memory, which is why precision is a latency knob before it is a memory knob. A full walkthrough of quantization and distillation, from what a number actually is, through outliers, SmoothQuant, GPTQ, AWQ and KV cache quantization, to soft labels, on-policy distillation, and the order you should apply them in.

By Codemia • Sep 25, 2026

28 min read

Coding agents now read a repository, plan a change, edit several files, run the tests, and iterate until they pass. A full walkthrough of the design, from requirements to sandboxing to evaluation, and what an interviewer is listening for at each step.

By Codemia • Aug 31, 2026

18 min read

Why QUIC’s user-space transport lets us ‘kill’ the old app-level event loop: a tour from Tahoe/Reno to Cubic/BBR, then into QUIC’s pacing, loss recovery, and stream-level multiplexing that trim tail latency and jitter at scale.

By Codemia • Oct 9, 2025

10 min read

From MySQL’s retired query cache to modern Redis/Memcached, this deep dive compares DB-level vs app-level query caching with real-world hit/stale rates and practical guidance.

By Codemia • Oct 4, 2025

5 min read