Case study
LMaaS
Self-hosted Apple Silicon model service: OpenAI-compatible API for local LLMs, plus ML train/serve/retrain behind SemVer releases.
- Go
- React
- TypeScript
Problem
My own apps needed local language-model inference and versioned ML releases without depending on a cloud provider for every call.
Approach
LMaaS is a self-hosted service on Apple Silicon: a Go API that speaks an OpenAI-compatible surface for local models, plus train, serve and retrain flows behind SemVer releases, with a React console.
Result
A private model desk my products can call locally, with a console demo live and source unpublished.