Data Scientist / ML Engineer
Building production ML systems for document analysis — OCR, NLP, and Document AI. End-to-end pipelines from model training to FastAPI services and Docker deployment. 3+ years in ML, 6+ years in Python.
Parsing structured data into Neo4j graph database, then building a GraphRAG web UI where users ask questions in natural language — an LLM generates Cypher queries and returns answers grounded in the knowledge graph.
Railway track segmentation using U-Net. Dataset of 89 labeled images expanded with augmentation to 984. Achieved IoU 0.80 on the test set.