Profile
About
I study Computer Science at VIT Vellore and build machine learning systems that have to work outside a notebook: offline on a phone, in a browser, or inside a library other people depend on.
Most of my work is computer vision and applied ML. I built BlindAssist, an offline navigation aid that decides what is worth saying to a visually impaired user, and ranked in the top 4% of the Amazon ML Challenge with an entity-resolution pipeline over ten million records. I presented a first-author paper on electricity demand forecasting at SIPCOV-2026, have a second, on BlindAssist, under revision at ICMLDE, and have two published patent applications with seven more filed.
I also fix bugs upstream, in kornia, timm and Haystack. The models on this site are trained from scratch, and every number on them cites the run that produced it.
Education
Vellore Institute of Technology
B.Tech, Computer Science and Engineering
Awards and certifications
Amazon ML Challenge 2026
Top 4% nationally, rank 696 of 89,000+ participants
Code for Life: Hackathon for Everyday Solutions, VIT
Winner, Best Project Award, Feb 2026
Google Cloud Certified: Generative AI Leader
Jul 2026
Extensive & Reimagined AI (ERA-V3)
Certificate of participation, 2024–25
Skills
- Languages
- ML / DL
- Engineering