Projects are where ideas become systems. Some are live, some are still experiments, and all of them document what I learned by building.
Designing and evolving the platform you’re using right now: a place for projects, AI experiments, technical insights, and documented learning.
A retrieval-augmented assistant for asking questions across private document collections while exploring chunking, embeddings, retrieval, prompting, and evaluation.
Testing chunking approaches, embedding strategies, similarity search, and retrieval quality across different document types and use cases.
Small focused experiments around prompt patterns, tool use, workflow design, agent behavior, and MCP.
New experiments, deeper case studies, and lessons learned will appear here as they move from idea to implementation.