Vijay Samant

During my internship, I had the opportunity to bridge the gap between the AI research and practical clinical application. I built and deployed a full-stack clinical decision support system designed to predict Multiple Myeloma staging using three distinct models: MRS, CRSS, and GCRS. The application was containerized with Docker and deployed directly to AIIMS servers for secure use by authorized medical professionals.

• AI/ML Integration: Implemented inference pipelines for a Graph Convolutional Network and pretrained decision-tree models. Processed ~700-dimensional clinical data vectors, including survival outcomes, ethnicity, and cytogenetic abnormalities.

• Backend & Database: Built a FastAPI backend with strict input validation and designed a PostgreSQL schema supporting role-based access control, audit logs, and secure patient data tracking.

• Frontend Development: Developed a React portal with dashboards for patient record search, single-patient analysis, and bulk CSV-based evaluations.