Startups · AI

Head of Data Engineering

Provision · San Francisco · On-site

← All jobs
About Provision

Provision reviews drawings, specs, and contracts in minutes. Catch risks earlier, avoid scope gaps, and bid with confidence. Purpose-built AI for general contractors and subcontractors in 2026. Backed by Y Combinator.

About the role

The head of the data engine must be someone with an intuitive understanding of what constitutes good data, as well as what types of data collection are required to push the frontier forward. They must have strong management and organization capabilities and have the ability to scale a team of programmatic and human annotators. Furthermore, they will be responsible for developing benchmarks and metrics to evaluate quality.

What they're looking for

  • A passion for data. This person should enjoy viewing data, and they bask in and immerse themselves in the particularities and idiosyncrasies that come from data distributions in the wild
  • A vision of what the future of the industry is regarding robotics data, including which types of modalities and distributions of data will be most effective as the sophistication of robotics develops
More about this role

Provision seeks to create a world of labor abundance through physical AI. Robotics will initiate a world of opulence in which the work we have long deferred finally gets done and services and tasks we never imagined become ordinary, at greater speed and scale. Our role in bringing this transformation into labor abundance is our edge in complex, scaled physical world services operations. That manifests as capturing immensely diverse, in the wild, high quality pre-training data, as well as every component of the stack in which there is a need for physical world complexity and a need for scale. This includes Physical world evals, multimodal collection, post training, as well as deployments.

Integrity. It is our firm conviction that trust and integrity are powerful economic forces, in their ability to reduce transaction costs and correct for information asymmetries.

Velocity of action. Acceleration creates a compounding competitive advantage.

Ownership. We believe in having an internal locus of control, and accepting responsibility for all outputs one has supervision and influence over.

The head of the data engine must be someone with an intuitive understanding of what constitutes...

Read the full posting on Provision's site ↗

Robotics

Build your edge while you search

Free tools for founders and investors, plus VC Unfiltered, our take on startups, venture and the people who build them.