Bagel Labs
Bagel Labs is an AI research lab building Distributed Diffusion Models, or DDM, which train frontier diffusion models across commodity hardware instead of one uniform GPU cluster by training many smaller expert models independently and combining them with a lightweight router at inference. We proved the approach with Paris, our first public model, and we are now bringing it to physical AI, where robotics, autonomy, and simulation teams need large models but cannot get the compute to train them the usual way.
We ignore years of experience and pedigree. If you can find the teams this matters to and turn a hard technical story into signed pilots, we want to hear from you. Every requirement below is flexible for the right person with enough judgment and network to open this market.
As our first commercial hire, you will find the physical AI teams whose training costs or hardware limits DDM can change, and move them from a first conversation toward a technical evaluation, a design partnership, or a paid pilot. This is how Bagel Labs sells into a market that barely exists yet, and how it gets done is up to you.
You have opened a hard technical market before, maybe as a founder, an early commercial hire, or the first business development or GTM person at an infrastructure or deep tech company. You know enough about ML infrastructure, robotics, or simulation to earn a technical buyer's trust, and you are commercial enough to turn that trust into a deal. You are comfortable working without a playbook, through a long sales cycle, on a product that is still taking shape. We care more about whether you can open this market than about your background.
Bagel Labs
Range