Full Job Description
About Delivery Ops
Delivery Ops turns raw captures into datasets customers can train on. We pull and filter episodes against each customer's spec, run every ship gate (quality, privacy, exclusivity, labels), package and ship the data, and report on what was delivered. As volumes grow into hundreds of thousands of hours per program, the team also owns showing the field what this data looks like and why it matters.
About the role
Every dataset we ship needs to be understood: by the customer receiving it and by the field deciding whether to work with us. You will write and publish the material that explains our data, from the documentation that ships with each delivery to the public technical posts that show what the data can do.
What you'll do
Delivery documentation
Own the dataset cards, data briefs, and documentation that ship with each delivery, written to a consistent standard.
Technical publishing
Write and publish technical posts, reports, and case studies on our datasets, methods, and results.
Public sample releases
Package and publish sample datasets and benchmarks with clear documentation and licensing.
Figures and analysis
Work with research and delivery to turn delivery data into accurate charts, examples, and explainers.
Review
Run every piece through technical, legal, and customer-confidentiality review before it goes out.
What you bring
• 3+ years of technical writing, science journalism, or developer documentation, ideally in robotics, ML, or computer vision.
• You read research papers comfortably and explain them accurately to both experts and buyers.
• Enough Python or SQL to pull your own numbers and make your own figures.
• A portfolio of published technical writing.
Even better if you have
• Research or engineering experience in robotics or ML.
• Experience publishing datasets or models on public platforms.
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.
