Predict.it — XGBoost training and deployment
Trained and optimized an XGBoost machine learning model using historical stock market data to predict future price ranges and market trends. Implemented automated data pipelines to collect, preprocess, and retrain models with updated data. The work also included deploying and managing model versions via Hugging Face and building an MLOps pipeline with GitHub Actions. • Built ML training workflow for XGBoost using time-series stock data • Implemented data collection and preprocessing pipelines for retraining • Integrated Hugging Face hosting for versioning and inference • Created GitHub Actions MLOps to retrain and deploy models