At AMLRightSource, I serve as MLRO, refine and conclude L2 investigations, and conduct STR/SAR filings with 100% quality. I use Chainalysis Reactor and OSINT workflows to trace multi-hop wallet transactions across blockchains.
As a freelance data scientist, I engineered and optimized more than 20 machine learning and deep learning architectures, improving baseline model accuracy by up to 30%. I also built preprocessing and model evaluation workflows for unstructured text and image data.
In my projects, I developed an RNN-based anti-money-laundering classifier and an audio emotion classifier using PyTorch. I also created a computer vision tool for analyzing chest X-ray images for cancer detection.
