DigiKey is partnering with engineering instructor and content creator Shawn Hymel for a free webinar on using reinforcement learning (RL) to train and deploy a self-balancing robot. The 90-minute virtual workshop, titled “Train a Balance Bot with Reinforcement Learning,” will take place on Thursday, 13 August 2026 at 10am CDT. It will guide engineers through training a robot in simulation before deploying the control policy onto physical hardware. Hymel will demonstrate using the M5Stack Bala-C self-balancing robot, showing how to import a 3D model into the MuJoCo physics simulator and train a neural network policy using a multi-phase Proximal Policy Optimisation (PPO) curriculum. The trained actor network will then be deployed to an ESP32 microcontroller using Arduino. Hymel noted that almost every major robotics lab uses reinforcement learning for robust control, but policy training can require millions of simulation steps and careful reward function development. Engineers are encouraged to use the M5Stack Bala-C Balance Bot or M5Stack Bala2 Fire Self-Balancing Robot. Those unable to attend live can register to receive the webinar recording. The webinar is accompanied by a six-part YouTube series covering simulation, reinforcement learning, and deployment to real hardware, with the first episode already available and a written tutorial published on DigiKey’s Maker.io platform. Registration is free through DigiKey.