At Toothlens Inc., I lead a team of five AI engineers across detection, segmentation, and generative AI. I architected a production vision-language pipeline that analyzes dental images, generates SOAP reports, and answers questions about patient scans across six countries.
I shipped SmartCheck, a dental-hygiene detection product serving more than one million users, and improved its DETR-based detection from 40% to 90% mAP@50. I also trained and deployed YOLOv12 for identifying missing, chipped, and misaligned teeth in patient scans.
I developed teeth-numbering segmentation and multiclass detection for dental conditions that became core vision technology for the Toothlens product. I also shipped Simmetry for cosmetic visualization and built a dataset-management and annotation platform that reduced annotation cycle time by 35%.
My work includes Brush-It, a beta AI brushing coach with real-time vision feedback, and research into diffusion-based dental visualization. I’ve also built a diffusion-based virtual try-on system and an X-ray workstation prototype.
