As a Freelance Data Scientist, I develop, train, and test machine learning models for client projects, including BERT-based NLP applications. I also created reproducible coding tasks and reference implementations for an ML infrastructure project, supported by automated Python tests.
I built Docker-based environments and validation workflows to check model predictions, training checkpoints, and reports, including edge-case testing for LoRA scaling.
At Zucol Group of Services, I automated internal data workflows with Python and built SQL transformation and validation pipelines. I also supported data migration, debugged scripts, and documented fixes.
For Sentinel, I built a Python and SQL platform combining analytical workflows with controlled LangGraph and RAG features. Its reports use validated calculations and citations, and I wrote 167 offline tests with 82%+ branch-aware coverage.

