VelocityInsight

Your personal AI race engineer—analyzing millions of telemetry points to optimize your racing line, predict strategy, and drive you to victory.

Comment

VelocityInsight: The AI Race Engineer

💡 Inspiration

Motorsport is the ultimate blend of human skill and engineering precision. In modern racing, data is king—but raw data is overwhelming. We were inspired by the complex telemetry screens seen on Formula 1 pit walls and wanted to democratize that level of insight.

Our goal was to answer a simple question: "Can we build an AI that doesn't just show you the data, but understands the race?" We wanted to create a tool that acts as a personal race engineer for sim racers and enthusiasts, turning millions of data points into a clear path to victory.

🧠 What We Learned

Building VelocityInsight taught us that context is everything. A lap time is just a number until you know the tire compound, the fuel load, and the track temperature.

  • We learned how to structure complex telemetry data so that an LLM (Google Gemini) could analyze it effectively.
  • We discovered that "premium" UI isn't just about colors—it's about motion. Smooth transitions and responsive interactions build trust in the tool.
  • We realized that in high-pressure environments like racing, clarity beats complexity. Every chart and icon needs to be instantly readable.

🛠️ How We Built It

We built VelocityInsight as a high-performance web application designed to feel like native software.

  • Frontend: We used React and TypeScript for type-safe reliability. The UI is styled with Tailwind CSS and features a custom "Glassmorphism" design system to mimic modern HUDs. We used Framer Motion for fluid animations and Recharts for high-speed data visualization.
  • The Brain: The core intelligence is powered by Google Gemini. We feed it race telemetry, weather conditions, and tire data, and it returns strategic advice—predicting pit windows and analyzing driving lines.
  • Visualization: We implemented custom WebGL shaders for our backgrounds and dynamic SVG manipulations for the interactive track map, which visualizes speed gradients in real-time.
  • Backend: A FastAPI (Python) server handles the data stream, simulating the firehose of information a real race team receives.

🏔️ Challenges We Faced

  1. Visualizing Speed: Rendering thousands of telemetry points on a track map without crashing the browser was tough. We had to optimize our rendering logic to create a smooth, color-coded speed gradient that updates instantly.
  2. The "Uncanny Valley" of AI: Initially, the AI advice felt generic. We had to refine our prompts and data structures to ensure Gemini understood racing nuances—like the difference between "saving fuel" and "lift-and-coast."
  3. Design Consistency: As the project grew, our UI became fragmented. We undertook a massive refactor to unify our iconography (moving to Lucide React) and theming, ensuring every pixel looked like it belonged in a professional paddock.

Built With

Share this project:

Updates