Kai Lo

Trained and evaluated LLM outputs by crafting high-performance prompts and criteria. Designed and improved AI/ML projects and solved complex mathematical problems to train LLMs.

Co-developed a proprietary index to gauge interest in 1500+ sports-related entities across 10+ markets, involving large-scale web scraping, API integration, and optimization of a custom search algorithm. Built and trained stacked regression models (linear, logistic, XGBoost) achieving 90%+ accuracy to forecast subscriber counts, viewership, and event attendance for major clients including Entain and Formula 1. Developed a motorsport racing simulator and race predictive models, and advised on an NBA Elo rating system. Collaborated with Data Engineers to deploy ML workflows on AWS and Git; maintained a CI/CD pipeline for model deployment. Conducted daily data processing and feature engineering using AWS, SQL, Databricks, and PySpark; experience with Kedro, Scikit-learn, Statsmodels, and LLMs. Created scalable web scrapers using Selenium, Requests, and automation tools; deployed on AWS EC2 and Lambda.

Taught college-level English, Mathematics, Calculus, Statistics, and Computer Science to students; provided subject-specific coaching and problem-solving strategies.