AI/ML Engineer | Agent Systems, RAG & MLOps
I build and operate AI systems, from model development and evaluation to deployment and monitoring. My work focuses on LLM applications, retrieval-augmented generation (RAG), and prompt security, with additional experience deploying models on NVIDIA Jetson.
I currently operate two production AI platforms:
- Empathika — An AI platform for healthcare.
- FinTech GraphRAG — A chatbot using Neo4j and graph-based retrieval for financial fraud and corporate intelligence.
I author two Python tools available on PyPI:
- personakit — A library for defining role-based LLM agents, with support for multiple model providers, tool use, and observability. PyPI
- spyv — A prompt-security tool for discovering prompts in supported Python integrations, testing for vulnerabilities, and adding runtime checks. PyPI
- AI & machine learning: Python, PyTorch, Hugging Face, LLM fine-tuning and evaluation.
- LLM applications: LangGraph, CrewAI, RAG, GraphRAG, Neo4j, prompt-security testing.
- Deployment & operations: FastAPI, Docker, Kubernetes, MLflow, Langfuse, GitHub Actions.
- Edge AI: NVIDIA Jetson, CUDA, TensorRT, C++.
My research includes an IEEE publication on banking cybersecurity.
For professional inquiries or collaboration, reach me at contact.majidul.islam@gmail.com.