We are hiring a
Forward Deployed Data Delivery Engineer to operate on the frontier with customers: take messy, real-world capture data - much of it raw video - and turn it into beautiful, reliable, model-ready datasets, while owning the technical relationship end-to-end.
This is a senior, high-trust role with significant autonomy. You'll combine data engineering, hands-on analysis, and product judgment to deliver datasets customers can train and ship on - and to make our delivery systems more reliable every time you do.
What You'll Work OnCustomer Delivery & Technical Ownership- Own the end-to-end delivery of customer datasets: requirements, validation, iteration, final handoff.
- Be the technical point of contact: communicate clearly, set expectations, and close loops.
- Turn one-off customer needs into durable internal improvements - tooling, pipelines, and standards that make every future delivery faster and safer.
Data Systems & Pipelines- Build, debug, and harden data pipelines across ingestion, transformation, QA, and export.
- Work fluently across storage and database paradigms (SQL + NoSQL + object storage) and pick the right tool for the job.
- Establish reliable dataset "contracts": schemas, versioning, provenance, and reproducible builds - so every dataset has a clear source of truth.
Dataset Quality & Signal- Define and measure what makes a dataset good for a given task: coverage, diversity, balance, label fidelity, and fitness for the customer's model.
- Build quality scorecards and coverage/diversity reports that make dataset health legible to customers and internal teams.
- Query and slice large corpora to maximize customer fit - surface exactly the data that matches a target distribution, not just bulk volume.
- When the signal a customer needs is missing or weak in the raw video, diagnose it and partner with the perception/ML pipeline teams to extract or improve it upstream.
Who You AreRequired Background- Significant hands-on experience designing, building, debugging, and operating production software systems. You should be comfortable going deep with senior engineers on architecture, APIs, data flows, failure modes, reliability, and technical tradeoffs-not just integrating or demonstrating existing tools.
- Proven experience personally owning substantial technical projects from an ambiguous problem through architecture, implementation, deployment, and production support.
- Direct experience working with external customers or partners to understand technical requirements, translate them into solutions, and remain accountable through successful delivery. Internal ticket-based support alone is not sufficient.
- High agency and a strong generalist bent - comfortable moving across data, delivery process, and whatever else the problem needs, taking on problems beyond your defined scope without waiting to be asked.
Nice to Have- Working literacy in video understanding, embeddings, and encoders - enough to reason about what a dataset teaches a model and where signal is missing.
- Experience building data-quality, coverage, or diversity tooling.
- Background adjacent to ML, computer vision, or robotics data.
Why This Role- Own the customer-facing delivery loop for world-class robotics datasets.
- High autonomy, high trust, and direct impact on customer success and revenue.
- Work across the full stack of the problem: data, pipelines, analysis, and delivery quality.
- Sit at the exact point where raw, messy, real-world data becomes the thing that makes embodied-AI models work.
A Note on ApplyingStudies show women and candidates from underrepresented groups often only apply when they meet 100% of the listed qualifications, while others apply after meeting 60%. If you don't check every box above but believe you can do the job, we encourage you to apply - we're looking for capability and trajectory, not a perfect checklist match.
