I build computer vision pipelines
Machine learning that survives contact with real data, real hardware, and real users.
About
I'm a Computer Science & AI student at Purdue. My work sits where machine learning meets software engineering: training and fine-tuning models, then building the pipelines and infrastructure around them so they're actually usable.
Projects & research
Project Manager · 6-person team
A computer vision pipeline that finds every rally, tracks every player, and cuts a full match down to its best moments plus structured stats. I lead three subteams (detection and tracking, event logic, video infrastructure) and personally built the rally segmentation and player-tracking modules. The bar: catch 85%+ of rallies on footage it has never seen.
Undergraduate Researcher · Low Power Computer Vision
Real-time deepfake detection for video calls that doesn't just flag a fake but explains why, in plain text. It's part of research into efficient models that run on constrained hardware, so a detector doesn't need a data center behind it.
Undergraduate Researcher · Privacy-Preserving ML
Combined several machine learning defense techniques into a single privacy-preserving pipeline across PyTorch and TensorFlow, and dug through the defense literature to test whether they hold up on tougher datasets and contexts.
Experience
Skills
Contact
Computer vision, ML systems, research, or just a good idea: email is the fastest way to reach me.