๐ Data Science @ RVCE | ๐ค GenAI & Agentic AI Builder
Shipping practical AI products at the intersection of systems, agents & real-world impact
I build AI systems that solve real problems โ from agent orchestration and RAG pipelines to production-ready products running on my own hardware.
I'm a fourth-year Data Science student at RV College of Engineering who spends more time in terminals than textbooks. My work sits at the crossroads of LLM systems, backend engineering, and self-hosted infrastructure. I run a 24/7 Ubuntu home-lab server called Friday that hosts everything from vector stores and multi-agent frameworks to Nextcloud, Jellyfin, and this portfolio.
What I care about:
- ๐ง Agentic AI & LLM Systems โ multi-agent orchestration, RAG, tool use, structured outputs
- ๐ Product Engineering โ turning ideas into working, deployed products fast
- โก Systems & Infrastructure โ Docker, Nginx, Cloudflare Tunnels, Tailscale, local LLMs
- ๐ Strong Foundations โ deep DSA practice and algorithmic problem solving
|
Local-first multi-agent AI framework A modular agent system with Commander, Researcher, Coder, Analyst, and Writer personas โ built to run locally with Ollama and exposed through a Gateway. Focus: Agent orchestration โข Local LLMs โข Plugin architecture โ Explore Repository |
Clinical AI with memory, FastAPI & Qdrant LangGraph agents triage symptoms, fetch prescriptions, and summarize cases; uses Qdrant memory, PyTorch diagnostics, and ElevenLabs voice. Focus: LangGraph โข RAG โข Medical AI โข Voice synthesis โ Explore Repository |
|
RAG-backed terminal chatbot on my portfolio Self-hosted Vite SPA with a zsh-style terminal assistant that answers from a Qdrant index. Hardened against prompt injection and secret leakage. Focus: RAG โข Qdrant โข Guardrails โข Self-hosting โ Visit Live Site |
AI product for large-scale health screening Building and deploying a structured clinical-AI pipeline in collaboration with Rotary District 3191 for 10,000+ beneficiary health camps. Focus: Healthcare โข Scale โข Real-world deployment โ Read on LinkedIn |
|
Sustainability tracking platform Intelligent tools for building sustainable practices through tracking, temporal analytics, and data-driven insights. Focus: Impact tracking โข Data insights โข Sustainability โ Explore Repository |
Satellite burn-severity mapping Sentinel-2 based burn-severity analysis via Google Earth Engine and FastAPI, turning satellite imagery into actionable environmental data. Focus: Remote sensing โข Earth Engine โข FastAPI โ Explore Repository |
| Goal | Current Work |
|---|---|
| ๐ค Agentic AI Systems | Production-grade multi-agent architectures & orchestration |
| ๐๏ธ System Design | Scalable backend patterns & self-hosted infrastructure |
| ๐งฎ Algorithmic Excellence | Deep DSA practice & problem-solving mastery |
| ๐ฌ AI Research & Products | Frontier LLM capabilities, RAG, and emerging AI patterns |
๐ก Interested in collaboration?
I'm actively looking to collaborate on:
๐ค AI Systems โข ๐ Developer Tools โข ๐ผ Startup Products โข ๐ฌ Research & Innovation
Let's build something extraordinary together!
Made with ๐ by Amritesh Sahu โข hosted on friday