I design reliable AI systems end to end — agents, RAG and LLM evaluation — with a Data Science foundation.
I design agentic systems and RAG that run in production.
Classic ML, NLP and anomaly detection on real datasets.
Research interests: AI for Robotics · Multimodal AI · VLM/VLAs · low-resource languages
I build the AI features of a compliance product that reads documents on its own and writes audit-ready reports — a LangGraph supervisor with specialized sub-agents, a Weaviate RAG knowledge base, sovereignty and security (AES-256, zero-retention), and continuous evaluation on AWS.
I built a system that matches CVs to job offers and measured how fair and accurate it was — a hybrid Content-Based + RAG recommender (LangChain, Llama-3, FAISS, Sentence Transformers), an LLM evaluation suite on an annotated golden dataset, and Azure MLOps pipelines.
2021 to 2024.
Deep Learning Specialization, Neural Networks with PyTorch (IBM), TOEIC 955.
Stories that think about the world.
Questions before answers.
Language as material.
The language of proofs.