AI Local Models Lab

A workspace for running models locally: Ollama workflows, converting and quantizing Hugging Face checkpoints, and small retrieval-augmented pipelines that keep context private. Each experiment tracks the task, the context the model receives, what output is useful, and what still needs human review — keeping the work grounded in applied workflows instead of prompt trivia.

More detailed notes — architecture, decisions, tradeoffs, and outcomes — get published here as the project matures, alongside source code and live deployments.