Brendan Marx

Brendan Marx builds robotic arms from the electronics up. He orders the parts, 3D-prints the joints himself, assembles the arm, and runs experiments to train it to move on its own. It’s the kind of work he came to AI Fund to do: take a physical-AI idea, build it, and find out whether it holds up.

Marx studied industrial engineering and computer science at the University of Illinois Urbana-Champaign, where he focused on optimization — the math of finding the best way to get something done when real-world limits get in the way. That instinct for making a system run as well as it can is what pulled him toward robotics, where the same problem meets the physical world.

On AI Fund’s technical team, he works hands-on with physical AI projects. Getting a robot to work is harder than it sounds, and Marx focuses on the part that matters most: not just programming an arm to repeat a fixed set of moves, but training it to respond to what’s actually happening around it — the difference between a machine that runs a routine and one that can handle a situation it’s never seen.

He’s especially drawn to the unsolved part of the problem: “The reason robotics isn’t at the same level as AI is because there’s not enough data,” he says. Closing that gap, along with giving robots a real understanding of the physical world, is the frontier he’s working on. What he’s after is generalist robots that can handle whatever you put in front of them — do the dishes, make the bed, fold the laundry, take out the trash. “A robot can do whatever you want it to do,” he says.

Before AI Fund, he interned with DeepLearning.AI’s special projects team, building agentic workflows and using sentence-transformers to create vector embeddings. He’s a Six Sigma and Lean Green Belt.