Full-Stack & AI Systems Developer
🏆 SIH 2025 Grand Finalist | IEEE 1st Prize Winner | GSoC-Ready OSS Contributor
📍 Pune, India | 🎓 AI & Data Science @ SPPU
Building production-ready AI-integrated web applications with React, Node.js, and LLM pipelines (RAG, Gemini API). Focused on scalable backend architecture, real-time systems, and solving real-world problems through code.
- 🥇 Smart India Hackathon 2025 – Grand Finalist
- 🥇 IEEE ICRTAIDS 2025 – 1st Prize for AquaSafe AI Beach Safety System
- 🥈 RUTAG Logo Contest – 1st Runner-Up (PSA Govt. of India)
AI & Automation: Google Gemini API • LangChain • OpenAI • HuggingFace • N8N
📚 PARRS — Personalized Academic Retrieval & Revision System
RAG-powered academic assistant where students upload study materials (PDF, PPTX, DOCX) and get source-cited, exam-format answers grounded exclusively in their own content. Achieved 93.4% groundedness and 0.8% hallucination rate across 275 evaluation queries.
Impact:
- Hybrid dense + BM25 retrieval pipeline with cross-encoder re-ranking delivering 1.4s avg end-to-end latency
- 6 exam-specific answer modes (2-mark, 5-mark, long-answer, MCQ) with tuned prompts — no other tool targets Indian semester exam formats
- Self-built evaluation framework (LLM-as-judge + rule-based) producing Recall@5, NDCG, MRR, groundedness, and hallucination metrics
- Multi-layer caching (Redis + LRU) with automatic invalidation on document upload; eliminated redundant embedding calls at scale
- Multi-tenant Row-Level Security across all tables for per-user data isolation
Stack: FastAPI, Python, Next.js 16, TypeScript, PostgreSQL (pgvector), Gemini Embeddings, Groq (Llama 3), flashrank ONNX, Redis, Supabase, Docker, Vercel + Google Cloud Run
📊 DataTrack — AI Document Automation System
Smart India Hackathon 2025 – Grand Finalist | Team Lead (Full Stack + AI)
AI system automating document processing for Kochi Metro Rail Limited, cutting manual processing from 8 hours to 20 minutes.
Impact:
- OCR-to-classification pipeline processing 1000+ multilingual documents daily
- Gemini API integration achieving 92% accuracy on structured entity extraction
- Semantic search improving document retrieval accuracy by 40% vs keyword-based search
- RBAC-enabled dashboards supporting 50+ multi-tenant users
Stack: Next.js, FastAPI, PostgreSQL, Supabase, Gemini API, Pinecone, AWS S3
🌿 Canopy — AI Project Roadmap Generator
Figma Makeathon 2025 | Solo Project
AI system converting unstructured project briefs into structured roadmaps, serving 200+ active projects.
Impact:
- Reduced roadmap creation time by 70% through automated task extraction
- Real-time collaboration with sub-200ms sync enabling 10+ concurrent editors
- 95%+ uptime through caching strategies and optimized database policies
Stack: React, TypeScript, Supabase, PostgreSQL, Gemini API
⛰️ WildSpire — Adventure Booking Platform
Solo Project
Full-stack booking platform with JWT authentication, handling 500+ bookings with 99.5% uptime.
Impact:
- 15+ REST APIs supporting activity discovery and secure payment processing
- Mapbox integration enabling location-based search for 50+ activities across 10 cities
- Rate limiting and XSS protection reducing malicious requests by 85%
Stack: Node.js, Express, MongoDB, Mapbox, Cloudinary
Specialized in AI-integrated full-stack applications with focus on:
- Production ML Pipelines: RAG systems, vector search, LLM orchestration (Gemini, LangChain)
- Real-Time Systems: WebSockets, sub-200ms sync, event-driven architecture
- Scalable Backend: Microservices, Redis queues, Docker, API design
Open to: SDE Internships | Full-Stack Roles | AI/ML Engineering | Backend Systems | GSoC 2026
💼 Actively seeking Summer 2026 internships in Full-Stack, Backend, or AI/ML Engineering
Let's build something impactful together!