Blog
Writing & thoughts
Notes on research, ML engineering, and ideas I'm exploring.
2026-02-10
Why ML Models Should Fail Gracefully
Most evaluation frameworks optimize for peak accuracy. But what happens when inputs fall outside the training distribution? I've been thinking about failure-first metrics and what they reveal about model robustness.
2026-01-25
Liquid Neural Networks: First Impressions
After months of working with LNNs for ECG classification at IIT Hyderabad, here are my observations on their interpretability advantages and the trade-offs compared to traditional architectures.
2026-01-08
Building for 100 Languages: Lessons from CVPR
Our paper on evaluating LLMs across 100 culturally diverse languages was accepted at CVPR 2025. Here's what we learned about the blind spots in multilingual AI evaluation.