Matthew Lee

Built JobCall, a real-time RAG-based AI voice assistant using GPT-4o-mini to automate property maintenance calls for 100+ properties and 22K+ residents, reducing manual escalations and improving unit-matching precision.

Designed a Databricks-based ML system (Smart Match) using XGBoost and RAG-style retrieval to automate carrier selection across 20+ carriers, improving matching accuracy and delivering auditable recommendations via SHAP.

Developed a LightGBM-based recommendation engine and AWS-hosted inference pipeline to automate technician dispatch and job routing, enabling real-time feature generation and automated retraining.

Delivered AI and data products including a BERT-based talent-matching engine and supply-chain risk dashboards, improving workforce utilization and operational visibility during a digital transformation program.

Mentored students in applied machine learning, reviewing code and guiding end-to-end workflows to improve project design, feature engineering, and model evaluation.

Improved manufacturing efficiency and automation for thermoplastic composite production through PLC/HMI integration, equipment layout design, and process data analysis.

Completed a Master’s program in Data Science focusing on applied machine learning and data engineering.

Completed the Machine Learning Nanodegree focusing on applied machine learning techniques and project-based learning.

Completed the Machine Learning course on Coursera covering core algorithms and model evaluation techniques.

Earned a Bachelor’s degree in Mechanical Engineering with coursework relevant to engineering fundamentals and process optimization.