Retro

Python, Machine Learning, Docker, XGBoost, PyTorch, Streamlit, SHAP, UE4, C++, Java, UE Blueprints

Advanced PyTorch for deep learning models, production deployment strategies with Docker, model optimization techniques, and model interpretability using SHAP analysis and feature importance.

ML applications with Docker deployment and unit testing. Current focus areas include fraud detection using XGBoost + SHAP, customer churn prediction, and Japanese sentiment analysis using BERT models.