About me

I’m interested not only in models, but also in the process around them: setting up experiments, checking results, and preserving enough context to revisit the work later.

I use this site to organize projects, notes, and papers that are ready to be shared publicly.

Skills and tools

Research work

  • Experiment setup
  • Run context
  • Result checking
  • Reproduction planning

AI and ML

  • Vision-language-action models
  • Model evaluation
  • Efficient execution
  • Research tools

Software

  • Python
  • TypeScript
  • Static websites
  • Git workflows

How I work

  • I leave a clear note when I do not understand something yet.
  • I record experiment settings together with their results.
  • I keep failed attempts when the reason is useful.
  • I check automated results myself.