AI/ML developer. I build agentic systems and full-stack products, then spend the rest of the week explaining to people what "agentic" actually means.
Currently: deep in multi-agent orchestration (LangGraph, CrewAI) and running local inference on Ollama — mostly because sending every prompt to a cloud API feels like paying rent on my own thoughts.
Not portfolio pieces — production/experience-level work with:
- LLM-powered content pipelines — generation + research + scheduling, glued together with FastAPI/React and a task queue that fires while I sleep
- Multi-agent workflows — Reader → Writer → Critic → Image → Publisher chains, i.e. teaching five small AIs to argue with each other so I don't have to
- Full-stack deployment — Node/Express + vanilla JS, Dockerized and shipped to Vercel, because sometimes the old ways are just the fast ways
- Local inference tuning — quantized models, GPU offloading, and enough
nvidia-smistaring to qualify as a hobby
| Area | Tools |
|---|---|
| Languages | Python, Java, JavaScript |
| AI / ML | PyTorch, TensorFlow, scikit-learn, LangChain, OpenCV |
| Agent orchestration | LangGraph, CrewAI, Claude API |
| Backend | FastAPI, Flask, MySQL |
| Infra | Docker, Git, GitHub Actions, Ollama |
Trying to graduate from "trains models" to "trains models that don't get owned in production." If you're already in AI security or AISecOps and want to tell me I'm doing it wrong, my inbox is open and my ego is durable.
📫 bhavishyachaturvedi@gmail.com — real emails only, no cold outreach templates, I can smell them