"I don't just build AI, I build AI that ships. RAG pipelines, multi-agent systems, LLM deployments if it's intelligent and production-ready, I'm probably already building it."
- 🤖 Specializing in RAG pipelines, LLM deployment, multi-agent systems, and prompt engineering - the full stack of making AI actually useful
- 🔬 Currently doing AI research - benchmarking LLMs (LLaMA vs GPT-4), building FAISS/Chroma vector pipelines, and squeezing out accuracy gains through smart chunking strategies
- 🧪 Built and deployed ML models across domains - income prediction (86% acc), resume parsing, startup evaluation, contract analysis - basically wherever AI can replace boring manual work
- 🛠️ Go-to stack: Python · FastAPI · LangChain · TensorFlow · PyTorch · MongoDB - comfortable taking a model from notebook to production
- 🏆 SIH 2024 Winner · ₹1L SSIP Grant recipient - proven track record of building things that judges and committees actually believe in
- 🏛️ VP, Computer Society & Gaming Club - led teams, organized events, and yes, fixed other people's merge conflicts
| 🏢 Organization | 🎯 Role | 📅 Period | 🔑 Impact |
|---|---|---|---|
| IPR, Govt. of India | AI Research Intern | Feb 2026 – Present | +18% retrieval accuracy · FAISS/Chroma pipelines · 3 chunking strategies benchmarked |
| Edunet Foundation | AI/ML Intern | Jun – Jul 2025 | 86% accuracy income predictor (XGBoost) · 48K+ records · Streamlit deployment |
| India Space Academy | Data Science Intern | Jun – Jul 2025 | 10+ astronomical datasets · ISS real-time tracker · 95%+ spatial precision |
|
Intelligent RAG Study Companion A full-stack AI-powered learning assistant that analyzes PDFs, URLs, and notes to generate contextual answers, summaries, flashcards, and quizzes. Stack: ✅ JWT Auth · Vector RAG · Multi-chat support · Quiz & Flashcard Generator |
CrewAI-powered Business Intelligence Evaluates startup ideas across Market, Financial, and Strategic dimensions using 3 specialist AI agents with real-world data from SerpAPI. Stack: ✅ Market Analysis · Financial Projections · Viability Scoring |
|
Autonomous Legal Associate A multi-agent AI system that reads PDF contracts, enforces a Company Playbook, flags policy violations, auto-redlines non-compliant clauses, and drafts vendor negotiation emails — end to end. Stack: ✅ OCR Ingestion · Compliance Auditing · Automated Redlining · Email Drafting |
AI-Powered Structured Extraction A production-grade FastAPI service that processes PDF resumes using OCR + OpenAI to extract structured data, generate embeddings, detect experience level, and expose a clean REST API. Stack: ✅ PDF + OCR · Embedding Generation · Experience Detection · Dockerized |
Agent Orchestration · Model Optimization · RAG Pipelines · Prompt Engineering · Document Processing
Tools: Docling · PyPDF · PyMuPDF · n8n · Make.com · Google APIs · ADK · Kaggle · Google Colab
| 🎯 Achievement | 📋 Details |
|---|---|
| 🏆 SIH 2024 | 1st Rank — Smart India Hackathon Internal Round |
| 💰 SSIP Grant | ₹1,00,000 — IoT Smart Environmental Monitoring |
| 🔬 Research | +22% RAG accuracy @ IPR, Govt. of India |
| 🤖 ML Model | 86% accuracy income predictor on 48K+ records |
| 🏛️ Leadership | VP — Computer Society & Gaming Club (2024–25) |
| 📜 Certifications | Stanford ML · IBM Deep Learning · IBM AI Fundamentals |
→ I can sleep 20 hours straight - not lazy, just in power-saving mode.
→ I turn last-minute panic into coding marathons (and surprisingly good results).
→ I'm equally excited by cricket highlights and hyperparameter tuning.
→ Debugging is my cardio (because running actual miles sounds like a lot).
I'm always open to collaborating on interesting AI/ML projects, discussing research ideas, or just chatting about tech.
Reach out — let's build something intelligent together!
"If you scrolled this far, we're basically friends now." 😄