Hello, I'm Lisa Cheng

I'm the cofounder and CEO of Arithmancy, an embodied AI company solving the sim-to-real gap in robotics. Robots trained in simulation learn from perfect information — the simulator tells the policy exactly where every object is. Then they get deployed into real kitchens and warehouses where they have to act on whatever their sensors can reconstruct, which is never perfect. They train on certainty and perform on guesswork. We generate physics-grounded training data that exposes robot policies to realistic perception error before deployment, so the failures happen in simulation — where they're cheap — instead of in someone's home. I found my co-founder, Mustafa Khan, by cold-DMing him on Twitter. Three months later we had a pipeline running on a real robotic arm.