RL Playground

Train real AI agents directly on your phone. Play, experiment, and discover reinforcement learning

RL Playground is a one-of-a-kind interactive experience to explore reinforcement learning through play

Agents don’t fake learning here — they actually learn, locally, on your device. No cloud, no servers, no precomputed simulations. You can watch learning happen live, episode after episode.

Each level is based on a concrete scenario:

a drone learning when to jump over a gap,

an autonomous car learning to avoid pedestrians,

a Duck Hunt–inspired challenge focused on tracking and decision-making.

You can tweak learning parameters, train the agent, then test it and observe how its behavior evolves over time.

RL Playground is made for:

people who want to understand AI hands-on,

curious minds, students, developers,

and players looking for experimental, meaningful games.

No internet required. No prior ML knowledge needed. Just learning by doing.