rmsle

Progressive Ames house-price modeling with metric-aligned MLPs, feature engineering, CatBoost, and a TabPFN-dominant ensemble.

[경진대회] 대구 교통사고 피해 예측 AI 모델 개발

Longitudinal learning project for the Cultivating ML Agent using a public time-series competition as a measurable training environment. Not a customer product.

Machine learning portfolio project: predict NYC Yellow Cab trip duration from GPS, datetime, holidays, weather and OSRM routing data. Full pipeline in Jupyter with Gradient Boosting submission.

LightGBM forecasting pipeline for the Nedbank Transaction Volume Forecasting Challenge (Zindi). Deep temporal feature engineering on 18M-row banking panel, optimised for RMSLE. Streamlit leaderboard tracker included.

AI Intern Project @leaderg inc.

Validation RMSLE obtained: 0.21163 which is less than the RMSLE score (0.22909) that won the Kaggle Competition.

End-to-end regression pipeline predicting calories burned during exercise using XGBoost, custom feature engineering, RMSLE optimization, SHAP interpretation, and Streamlit deployment.