Inspiration
Mobility competitions like drone racing and autonomous driving are expensive and risky to test in real life. We wanted a virtual, safe, and interactive way to simulate and compare AI-driven agents under realistic conditions.
What it does
SimRace AI is a browser-based simulator where multiple agents (vehicles, drones, or bots) compete in real time. Each agent follows rule-based or AI-driven logic, and a live 3D dashboard shows positions, collisions, and leaderboard rankings.
How we plan to build it
Frontend: React + Three.js for 3D visualization Backend: Node.js + Socket.IO for real-time data flow Simulation: Python (FastAPI + NumPy) for agent logic Infra: Redis for telemetry and leaderboard updates Hosting: Vercel + Render for quick deployment
Challenges we expect
Ensuring smooth real-time synchronization between simulation and visualization, maintaining determinism for replays, and balancing performance with visual fidelity.
Accomplishments we aim for
A working demo where multiple agents compete visually in real time, accessible entirely through a browser.
What we’ll learn
Cross-language communication between Python, Node, and React; designing deterministic simulations; and managing live data pipelines.
What’s next
Integrate machine learning agents using PyTorch, expand environment physics, and create an online leaderboard for public competitions.
Log in or sign up for Devpost to join the conversation.