I build production-grade AI systems that solve real-world problems, specializing in Intelligent Document Processing and AI Compliance Systems.

• 🎯 28+ Public Repositories with live deployments
• 🚀 5+ Production AI Systems serving real users
• 💻 Tech Stack: Python, TypeScript, React, Next.js, FastAPI
• 🌟 Specialization: AI/ML Engineering, Full-Stack Development
• ☁️ Azure Certified (AZ-900)
• 📈 95%+ Accuracy on ML evaluation frameworks

• 🎯 AI/ML Systems: Building evaluation frameworks, NLI models, RAG architectures
• 📄 Document Intelligence: PDF to structured data with 95%+ extraction accuracy
• 🔒 AI Compliance: Hallucination prevention, drift detection, audit systems
• 🌐 Full Stack Development: Next.js + FastAPI production deployments
• 💰 Fintech/Offline-First: PWAs with sync engines, payment integration

AI compliance system for corporate sustainability reports

• Tech: Python, GPT-4o, Pydantic V2, NLI Cross-Encoders, RAG
• Impact: 92% agreement with human auditors, automated 40+ hour audits
• Highlights: Drift detection (0-3 scale), Sankey visualizations, zero hallucinations

LLM evaluation framework with multi-metric scoring

• Tech: Python, Sentence-Transformers, PyTorch, HuggingFace
• Impact: 95% test coverage, 34% drift detection across test datasets
• Highlights: Relevance, groundedness, toxicity metrics with caching optimization

AI document processor converting PDFs to structured Excel

• Tech: React, FastAPI, MongoDB, OpenAI, WebSockets
• Impact: 95%+ extraction accuracy, processed 1000+ documents
• Highlights: Real-time streaming, self-improving AI system

Offline-first financial PWA with bi-directional sync

• Tech: Next.js 15, FastAPI, PostgreSQL, IndexedDB
• Impact: Zero-latency UX in low-connectivity areas
• Highlights: Custom sync engine, conflict resolution, 95+ Lighthouse score

Perplexity-style RAG chat with precise citations

• Tech: Next.js 14, FastAPI, PostgreSQL, LangChain, FAISS, SSE
• Features: Real-time streaming, vector search, queue-based architecture
• Highlights: Page-level citations, animated generative UI

Enterprise backend with async task processing

• Tech: Django 4.2, PostgreSQL, Celery, Redis, Docker
• Features: Async task queues, Excel data ingestion, Swagger docs
• Highlights: Containerized deployment, comprehensive testing

Production platform with payment gateway

• Tech: Flask, MongoDB, Razorpay, OAuth, JWT
• Features: Payment processing, PDF receipts, role-based auth
• Highlights: Live production system serving real users

Agentic meeting analyzer with autonomous workflows - Tech: Next.js 14, FastAPI, LangGraph, RAG (FAISS/pgvector) - Highlight: "Built a Citation Engine that highlights exact source paragraphs in PDFs. Used LangGraph to manage the chat state and SSE for real-time token streaming."

• Talk-to-Figma (MCP)

  • Tech: TypeScript, Model Context Protocol
  • Highlight: "Optimized the MCP Server for Figma to enable npx execution on Windows, resolving cross-platform compatibility for AI Agent workflows."

• Insight Engine

  • Tech: Python, OpenAI, Custom Agents
  • Highlight: "Architected a Multi-Agent Router system from scratch to dynamically select tools for meeting analysis. Currently migrating logic to LangGraph."

AI-powered educational platform

• Tech: React, Vite, Tailwind, OpenAI
• Features: Document processing, AI summaries, interactive Q&A

• Data Science Internships/Roles
• AI/ML Engineering Positions
• Full-Stack Development (AI-focused)
• Research Collaborations in NLP/Document Intelligence