Hi, I'm Atharva. I am interested in machine learning, specifically: interpretability and physical AI, and open-source software.

• ML research:
  1. tabfm-interp, tabular foundation model interpretability (paper: arxiv.org/abs/2606.12917),
  2. and a SemEval 2026 submission (paper, POLAR-SemEval-Submission)
  3. hierarchical transformer for detecting adversarial packet capture data: https://github.com/dorkcubed/sparkle, deployed over FastAPI and Docker
• ML tooling: ml-experiment-navigator, a tool for navigating/tracking ML experiments
• Robotics & simulation: arena-robot-simulation (line-following robot) and mec-simulation (Mecanum robot)
• Web projects: anant-website, Maths-Assoc-Website
• Open-source contributions: PyTorch's ignite, and user-scanner (a cross-platform username-availability checker)

Also check out: https://standardebooks.org/ebooks/anne-bronte/agnes-grey