In my Computer Vision Research Intern work, I evaluated YOLO, MediaPipe, and Keypoint R-CNN for human keypoint detection under varying video quality and degradation conditions.
I built a video preprocessing and evaluation pipeline that tested model robustness by reducing frame rate and resolution and introducing Gaussian noise. I compared predicted and ground-truth keypoints using Euclidean Distance, MAE, RMSE, MJPE, and PCK.
I also engineered AI Quiz Generator, a full-stack assessment platform that creates custom quizzes from PDFs and text using the Gemini API. I modeled its relational data in PostgreSQL with Prisma and implemented authenticated REST API routes and an analytics dashboard.
I'm studying Computer Science Engineering at Vellore Institute of Technology, Chennai. My technical work spans machine learning, computer vision, full-stack development, and database-backed applications.
