I am working on something new at Vmax. We are building agents that leverage the inherent structure in large company datasets to infer trajectories of states, actions and rewards. With this data, we are building multistep RL agents to carry out long-horizon tasks. See our preview!

vmax-ai.com