Building in Public

Technical work rarely moves in a straight line. A finished repository or paper shows the result, but it hides the experiments, false starts, tradeoffs, and small discoveries that made the result possible.

I am rebuilding this site as a public working record. It will collect three kinds of material:

  • Projects — research software, AI systems, data pipelines, and prototypes.
  • Learning notes — explanations and experiments from my ongoing study of deep learning, algorithms, and agentic systems.
  • Research — publications and practical lessons from computational genomics and epigenomics.

My goal is not to publish a stream of polished announcements. I want to leave a useful trail: what I tried, why I chose one approach over another, what broke, and what I would change the next time.

The first step is removing the template and making the site honestly mine. The next step is doing the same for the work it documents.