I build end-to-end AI systems — Next.js frontends, FastAPI backends, and the LLM/agent layer that ties them together. B.Tech CSE, UPES (2022–2026), with an IBM internship and merged open-source work. Based in Alwar, Rajasthan. Open to Full-Stack AI Engineer roles.
- 🛒 SnapKart — WhatsApp AI ordering agent for kirana stores, shipped as a merged PR (#252) to
Lamatic/AgentKit - 🎫 Resolvix AI — 5-microservice support platform with RAG triage, Redis Pub/Sub, and Groq
- 🎙️ Voice Shipment Tracker — real-time voice agent on LangGraph + MCP + WebSockets
- 🔧 Going deeper on event-driven backends: RabbitMQ, pgvector hybrid search, Postgres RLS
| Languages | Python · TypeScript · SQL · C++ |
| GenAI | LLMs · RAG · LangChain · LangGraph · Multi-Agent Systems · MCP · Prompt Engineering |
| Models & APIs | Groq · Gemini · Hugging Face · Whisper |
| AI Infrastructure | pgvector · FAISS · ChromaDB · Embeddings · Hybrid & Semantic Search |
| Backend | FastAPI · REST APIs · WebSockets · Redis · RabbitMQ · Microservices |
| Frontend | Next.js · React · Tailwind |
| Databases | PostgreSQL (RLS, pgvector) · MongoDB · SQLite |
| Cloud & DevOps | Docker · AWS (EC2, S3) · Render · Vercel · Git · GitHub Actions |
🛒 SnapKart · merged into Lamatic/AgentKitWhatsApp AI ordering agent for Indian kirana stores. Intent classification plus hybrid search (Gemini embeddings + pgvector) for product retrieval, wired end-to-end across Twilio, Airtable, and Slack.
|
Customer-support platform across 5 microservices — API gateway, auth, tickets, AI triage, notifications. Event-driven via Redis Pub/Sub, with a RAG triage pipeline (Chroma + Groq) surfacing classifications and response suggestions in a Next.js dashboard.
|
|
Real-time voice agent over WebSockets — Whisper transcription, LangGraph intent routing with Groq, and shipment lookups exposed through a custom MCP tool server. Async TTS, containerized for concurrent queries.
|
☁️ EcoCloud · IBM Summer InternshipCloud data-center simulation platform built in a team of five. ML-driven task scheduler improved server utilization ~25–30%, with a live dashboard tracking performance and energy usage.
|
More projects
Multi-Format RAG Chatbot — Chat with PDFs, DOCX, CSVs, and websites. FastAPI backend, FAISS retrieval, Streamlit UI. FastAPI LangChain FAISS Gemini
AI Crypto Forecast — ML-driven forecasting applied to crypto market data. Python
