Generalist Data Partnerships role at Generalist owning the data engine that feeds embodied foundation models. Responsible for finding and managing data partners globally to supply quality and volume data for robotics models.
Responsibilities
Embodied foundation models and their capabilities are driven by data — and robotics data does not exist on the internet waiting to be scraped. It has to be manufactured by people doing real work in real environments, to a spec guided by our model science and scaling laws.
The hardest part of the job is translating model needs and data requirements into something a data partner can execute without you in the room (and you are the first to catch the drift when what comes back is technically compliant but practically useless).
You’ll be responsible for:
Data supply strategy. Identify what we need to buy versus build. Map the vendor landscape across data types and quality. Bring recommendations, not options.
Vendor sourcing and diligence. Find, evaluate, and pilot new data partners. Run structured trials before committing volume. Know how to tell a real capability from a good deck.
Qualification
Translation and specificationCommercial termsScaling a partnership into a networkOperating cadenceClosing the loopTravelYou might thrive in this role if youExceptional written communicationBackground in autonomous vehicles
Required
Translation and specification. Author partner-facing specs that a non-ML operator can execute against with no follow-up call. Define the unit of delivery, what passes, what fails, and what to do when it is ambiguous.
Commercial terms. Structure pricing, rate cards, minimums, milestones, and acceptance language. Work with legal on MSAs and SOWs. Own the unit economics and know what we are paying per unit of usable data, not per unit delivered.
Scaling a partnership into a network. When one partner hits capacity, stand up the next without dropping quality or blowing up cost.
Operating cadence. Trackers, weekly partner reviews, forecasts against the research roadmap, and a clear picture of cost, volume, and quality that anyone at the company can read.
Closing the loop. Report back to research on what the data actually produced and use it to change the next cycle's spec.
Travel. Periodic domestic and international travel to partner sites and collection operations.
You might thrive in this role if you:
6+ years owning external partners or vendors who delivered a product or service to you — supply chain, vendor management, outsourcing/BPO management, data operations, or partnerships with a delivery obligation.
Demonstrated experience standing up a new vendor from zero: selection, diligence, contracting, ramp, and ongoing management.
A track record of translating technical or specialist requirements into unambiguous instructions that a non-technical execution partner delivered against successfully.
Fluency with commercial mechanics — pricing structures, SOWs, acceptance criteria, unit economics — and the judgment to know which terms actually protect quality.
Exceptional written communication. You will write documents that people you have never met execute without you present.
Preferred
Experience buying or managing data collection, annotation, or labeling at scale — managed workforce, crowd, or BPO.
Experience close to an ML or research organization, and enough working understanding of training data to push back on a request rather than just relay it.
Background in autonomous vehicles, robotics, hardware, or another domain where physical-world data acquisition is a first-class problem.
Experience at a company that grew quickly enough that the process you inherited stopped working and you had to rebuild it.