Data Scientist
Current• Developed manufacturing schedule optimiser app using Gurobi, Julia, and Python to replace the existing suboptimal manual process. The optimiser is planned to be used daily over 28 sites across New Zealand helping to process about 16 billion litres of milk everyday. The app is estimated to save the company $3‑10M annually.• Developed energy consumption prediction model using random forest regressor, Scikit‑Learn, Databricks, and MLflow. This resulted in saving the company $218K annually.• Implemented an incremental-loading logic for a large output table which was being fully-loaded daily. My implementation reduced the daily computing time from over 90+ minutes to consistent 15 minutes which allowed the reporting to be made efficiently at a critical time for the stakeholder. The Databricks notebooks I developed for this project is now the standard template across the team for similar processes.
