← All projects

AI Engineering

ML systems & inference

Implementation-led study of the systems choices behind useful, reliable LLM applications.

StatusOngoing technical work
Focusinference · serving · reliability

Why this exists

An ongoing engineering area spanning inference paths, KV caching, attention variants, quantization, batching, GPU memory, serving, observability, and production Python architecture. It is presented as technical work in progress—not as a claim of a particular deployment scale.

Working notes

This page is intentionally a concise record rather than a polished case study. As the work matures, it can hold architecture diagrams, experiments, implementation notes, and links to associated writing.