I'm a CS researcher who genuinely enjoys building things that matter — whether that's a mobile app helping elderly patients get oral health screenings, a system tracking orthodontic progress remotely, or models helping scientists discover new materials faster.
I just wrapped up an accelerated 3-year BS in Computer Science at the University at Buffalo (yes, 3 years — it was a lot of coffee ☕), and I'm heading to Michigan State University this fall to start my PhD in the Data Mining Lab under Dr. Pang-Ning Tan. I'll be working on spatiotemporal machine learning, deep learning-based weather forecasting, and AI adversarial robustness.
The thing I find most interesting is the gap between a model that works in a notebook and a system that works in the real world. I've spent the last couple of years living in that gap — and I wouldn't have it any other way.
Check out everything I'm up to at dkritarth.com ✨
|
🦷 Mobile Health AI @ ESC Group, UB Built OralScan — a mobile app with YOLOv8 vision pipelines for geriatric dental screening. Then OrthoScan for remote orthodontic monitoring using 3D GNNs. Currently working on mRehab — a telerehab platform with AWS backend and Kalman filter-based activity detection from wearables. Co-authored a paper submitted to Smart Health. |
⚗️ Materials ML @ Peng Lab, UB Benchmarked symmetry-aware GNNs on perovskite ordering energetics, and implementing E(3)-equivariant interatomic potentials with alchemical graph extensions. Co-authored a ChemRxiv paper on agentic AI for catalyst discovery. |
| 🏎️ PaddockPsychRL | 🛸 Marine Guardian | 🦷 OralScan |
|---|---|---|
| Multi-agent RL where real F1 lap-time data shapes agent psychology. MAPPO psych-conditioned policies improved Strategy Resilience Score by +16.7% | Ship detection in satellite imagery — 98.72% accuracy using a geometry-first pipeline before going deep | AI dental screening app — 2nd place at UB Health Futures Challenge |
I did my undergrad in 3 years (21–22 credits a semester) while TAing, tutoring, and running two research positions simultaneously. Somehow I also found time to build a side project about Formula 1 driver psychology. I think the best researchers are the ones who can't stop being curious about things outside their main field — and I'm working hard to stay that way.
If you're working on something at the intersection of spatiotemporal ML, scientific ML, or AI for health — I'd love to chat. My inbox is always open 🙂
