About
Full-Stack Developer (CGPA 8.99) with three production internships and three independently deployed full-stack systems across the MERN and Next.js stacks. Highlights include a 40+ endpoint REST API healthcare backend, an AI-driven task-triage agent built on Claude Haiku, and a CNN achieving ~90% accuracy for real-time oil-spill detection. Combines strong software engineering fundamentals with applied AI/ML integration and Jest-based testing practice.
Experience3
Full-Stack Development Intern
Dec 2025 — Mar 2026MediGhar Healthcare Services
Architected healthcare backend with 40+ REST API endpoints for auth, e-commerce, and healthcare, including real-time doctor-patient chat using Socket.io with JWT authentication and presence tracking. Secured platform using JWT auth, bcrypt hashing, Helmet.js, CORS, rate limiting, and input validation across all endpoints.
Web Development Intern
Dec 2025 — Feb 2026Olcademy
Built district-level informational card components in React and Tailwind CSS with Jest unit tests; independently shipped features in Agile sprints. Diagnosed and fixed recurring notification icon bugs and improved navigation/UX flow, eliminating UI defect classes.
AI Intern
Sep 2025 — Nov 2025Infosys Springboard 6.0 (Project: DeepSpill)
Trained and deployed CNN using TensorFlow/Keras achieving ~90% accuracy for real-time oil-spill detection in satellite imagery, reducing false positives using OpenCV filtering. Delivered dual web interfaces (Streamlit dashboard and React/FastAPI stack) with automated email alerts and SQLite-based detection logging.
Projects3
Engineered an AI Triage Agent using Anthropic Claude Haiku with 4 tool-calling functions to propose task priority, owner, and SLA, auto-applying at 75%+ confidence with human-in-the-loop fallback. Built an NLP classification pipeline with auto-retraining, 5-fold cross-validation, and rollback safeguards; designed an interactive React Flow dependency graph with custom bottleneck detection.
- Anthropic Claude Haiku
- Naive Bayes NLP
- React Flow
Developed a full-stack booking platform with Razorpay payments, automated host commission splits, and a dynamic pricing engine serving real-time availability. Integrated Google Gemini AI for conversational search chatbot and itinerary generation; implemented OAuth 2.0 authentication and Redis caching. Containerized platform with Docker Compose and Kubernetes.
- Razorpay
- Google Gemini AI
- OAuth 2.0
- Redis
- Docker Compose
- Kubernetes
Engineered browser-based face recognition using face-api.js with 128-dimensional descriptor matching, multi-angle enrollment, and Euclidean distance scoring. Built a real-time webcam pipeline processing frames every 300ms achieving ~95% accuracy under normal lighting. Implemented offline-first architecture with Dexie.js/IndexedDB and MongoDB backend sync.
- face-api.js
- TensorFlow.js
- Dexie.js
- IndexedDB
- MongoDB
Education1
Bachelor of Engineering
2023 — 2027Atharva College of Engineering, Mumbai, Maharashtra
Information Technology