Cartpole

Building reinforcement learning environments

Y CombinatorSpring 2025ActiveB2BReinforcement LearningData LabelingAIML

What they do

Problem

Frontier AI models lack high-quality reinforcement learning environments needed for effective training and evaluation.

Solution

Cartpole builds reinforcement learning environments used to train and benchmark frontier models.

Who it's for

AI researchers and developers training and evaluating frontier machine learning models.

Mateo Perez

Founder

Building Cartpole. PhD @ CU Boulder.