GridLM

A system that utilizes large language models (LLMs) to provide personalized coaching for Forza Motorsport players by analyzing their gameplay telemetry data like speed and throttle. This innovative approach aims to bridge the skill gap between casual and competitive racers by transforming complex game data into actionable performance insights. GridLM employs a decoupled architecture, using telemetry data processed and stored before being analyzed by LLMs integrated with LangChain to generate tailored recommendations. Evaluations of the system focused on feedback coherence, token efficiency, and hallucination reduction, demonstrating its potential as a valuable tool for gameplay improvement.