race-strategy

My personal project for F1 race strategy analysis and visualisations, including "What made the difference?" reports/blogs.

Open-source multi-agent AI platform for Formula 1 race replay, strategy simulation, and explainable race strategy recommendations. (work in progress, MVP in legacy branch)

Open-source race engineer simulator driven by real-time vehicle dynamics. Drive the strategy, not the car. AI executes. Physics punishes.

An F1 22 telemetry parsing web-app integrating a voice-driven AI race engineer, built to interpret live telemetry, simulate strategy calls, and deliver real-time strategic feedback. (WIP. INCOMPLETE)

Monte Carlo simulator that ranks F1 two-stop tyre strategies using real FastF1 telemetry, modelling tyre degradation, fuel burn, and safety-car probability over 10,000 race sims.

Production-ready F1 strategy simulator with real-time telemetry analysis, pit optimization, and faster API responses via Redis caching.

F1 undercut strategy simulator built on real FastF1 telemetry. Deterministic pace model validated against 2025 Hungarian GP — 100% track-position agreement, ~0.15s/lap pace error.

My project for F1 race strategy analysis and visualisations

Local-first endurance race strategy agent with deterministic stint and fuel tools, telemetry auditing, and optional Ollama.

Real-time F1 telemetry & race strategy engine that decomposes dirty air, fuel burn, and track grip to isolate true tire degradation and predict pit-stop windows with Gemini AI.

Always-on Race Intelligence System

A live Formula 1 race strategy engine — ingests F1's live timing feed, models tyre degradation and safety cars, and simulates the remaining race to answer: should we pit now?

Production-grade Formula 1 telemetry dashboard built with React and Vite for multi-driver lap analysis, tyre degradation modelling, race strategy simulation, and high-performance custom SVG data visualizations.This works well because it highlights: production-grade architecture domain depth (F1 telemetry) performance engineering data visualization

Hybrid deterministic + multi-agent Formula 1 pit-wall strategy system.

Physics-based and optimisation-driven modelling of cycling breakaway strategies for different rider archetypes under varying race conditions.

AI-powered community setup research and race strategy app for F1 25

A Formula 1 race strategy analyst that pulls real timing data, remembers past analyses, and explains pit stop decisions, tyre strategies, and driver comparisons with confidence scores.

Local-first, open-source race engineering companion for Le Mans Ultimate.

AI-powered F1 race strategy copilot using IBM Granite + Docling + FastF1 telemetry + Q-Learning RL agent. IBM SkillsBuild AI Builders Challenge May 2026.

Offline, evidence-grounded research toolkit for F1 race-strategy decisions: Bayesian pace modelling, Monte Carlo simulation, constrained optimisation, and cited LLM explanations.