RaceSense

• 

  hero

RaceSense — Hackathon Submission ✨ Inspiration

Grand Prix racing produces massive amounts of data, but race engineers often have only seconds to interpret it and make critical decisions. We were inspired by the idea of creating an AI co-pilot—something that can think faster, analyze deeper, and predict smarter than any human team alone. RaceSense was born from the desire to give racing teams Formula-1-level intelligence, even if they don’t have Formula-1-level budgets.

🏁 What it does

RaceSense is an AI-powered race strategy and telemetry intelligence app that helps racing teams win through data-driven decisions.

It provides:

Real-time telemetry monitoring (speed, tyre temps, fuel load, ERS, GPS)

AI strategy predictions (optimal pit windows, undercut/overcut options)

Tyre wear forecasting using ML

Driver performance analytics with braking/acceleration insights

Weather & track condition prediction

Team collaboration chat + instant alerts

Fan mode with simplified dashboards and AR race overlays

RaceSense turns complex datapoints into action: pit now, push harder, change tyres, avoid overheating, prepare for rain, etc.

🛠️ How we built it

RaceSense was built through a combination of:

Tech Stack

Frontend: Flutter/React Native

Backend: FastAPI + TimescaleDB for telemetry

AI Models:

LSTM models for tyre degradation

Gradient boosting for overtaking probability

Weather prediction via external APIs

Vector-based driver performance comparison

Visualization:

Real-time charts

GPS heatmaps

Live strategy simulations

Development Process

Started with wireframes for high-speed UI

Built real-time ingestion pipelines

Trained prediction models on sample racing datasets

Simulated races to validate pit strategy logic

Added collaboration and fan experience modules

🚧 Challenges we ran into

Handling real-time data ingestion without lag

Balancing accuracy vs. speed of AI models

Designing a UI that works under pressure (race engineers think in milliseconds!)

Getting telemetry datasets that matched real motorsport environments

Building simulations that felt realistic and not random

Integrating weather prediction with strategy logic

🏆 Accomplishments that we're proud of

Built a functioning AI strategy engine that predicts optimal pit stops

Achieved consistent tyre wear forecasting using ML

Created a smooth, race-ready dashboard with zero stutter

Delivered a fan experience mode that brings racing analytics to the public

Integrated driver performance comparison that identifies micro-gains

Our biggest achievement: turning complex motorsport analytics into a simple mobile app

📚 What we learned

Real-time data systems require different architecture than traditional apps

Racing strategy is a blend of science, prediction, and intuition

UI/UX for high-speed environments must be minimal, bold, and instantly readable

Building simulations reveals how small decisions change entire race outcomes

AI isn’t just a tool—it can be a teammate

🚀 What’s next for RaceSense

Integrating with real racing telemetry providers

Deploying RaceSense for amateur and professional racing teams

Adding voice control for drivers (“AI, what’s my tyre wear?”)

Building 3D race reconstruction using AR

Expanding to karting, NASCAR, MotoGP, and EV racing

Launching a premium analytics subscription

Eventually becoming the standard race engineer assistant for motorsport teams worldwide

Built With

• amazon-web-services
• and
• built-with-flutter
• deployed
• docker
• fastapi
• on
• postgresql/timescaledb
• python-ml-models-(tensorflow/xgboost)
• real-time-node.js-streaming
• with