Ojasvi Saini

Built an end-to-end machine learning pipeline covering ingestion, validation, transformation, model training, evaluation, and deployment. Developed and deployed a FastAPI-based REST service using Docker, MongoDB, AWS S3 for model registry, and AWS ECR. Implemented CI/CD with GitHub Actions including model versioning, artifact tracking, and automated deployment. Evaluated models using Champion/Challenger strategies with SMOTEENN balancing and comprehensive logging to support reliable iteration and performance comparison.

Developed an AI-assisted healthcare platform using the MERN stack for patient record management, doctor appointment booking, and AI-powered medical assistance. Built secure patient and doctor portals with JWT authentication, Cloudinary document storage, appointment scheduling, and health record workflows. Implemented an asynchronous AI pipeline where uploaded documents are stored in Cloudinary, indexed in MongoDB, and processed by a dedicated FastAPI service to create OCR-based structured summaries. Designed a dual-context RAG chatbot using LangChain to support both general medical conversations and patient-specific answers grounded in extracted records. Built REST APIs using Node.js, Express.js, FastAPI, and MongoDB to integrate authentication, OCR processing, AI inference, conversation history, and appointment features.