Postdoc. See my papers here.
Trying to understand generative modeling, reinforcement learning and decentralized finance.
Сontribution:
My main theoretical impact is a contribution to the understanding of optimal transport in machine learning. Especially rethinking optimal transport for reinforcement learning.
From an engineering standpoint, I helped design a neural computer architectures that generated molecules with favorable pharmacokinetics. These were later verified in the real world.
From a purely practical perspective, I developed a cowswap solver and various AI layers for executing swaps, bridging assets, and managing decentralized finance.
To share my knowledge, I designed a few courses on reinforcement learning and deep generative models, from GANs to flow-matching.
If you're interested in collaboration or discussing, I'd love to connect.
Social: @machinestein
Last update: June 11, 2025