The Role
We are hiring a Software Developer to build - end to end - the internal products and tooling that turn Mecka's ego-data pipeline into fast, reliable, at-spec dataset deliveries to frontier AI labs. This is a 01 role: you'll take a rough delivery problem, design the interface and the system behind it, ship a first version fast, and iterate with the people using it.
You'll own whole surfaces rather than a single layer - the review and QA UIs our labeling and delivery teams live in, the dashboards sales uses to track a delivery, the export and validation logic underneath, and the data plumbing that ties it together. Front-end one week, data pipeline the next, a scrappy internal tool the week after.
You'll work hand-in-hand with the Delivery Operations team - including our Robotic Data & Deliveries TPM - turning recurring delivery pain into products people actually want to use. When a workflow is slow, manual, or fragile, you're the one who builds the thing that fixes it.
Ideal background: full-stack or product engineering with a track record of building tools and products 01. Comfort with data and large-dataset systems is a strong plus, but breadth, product sense, and speed matter more than depth in any one stack.
What you will be doing:
Internal Products & Delivery Tooling
• Design and build the internal products our teams use every day - dataset review and QA interfaces, delivery-tracking dashboards, data viewers, and the workflows that string them together - owning both the UI and the logic behind it.
• Turn one-off scripts and manual steps into real tools with a usable interface, so delivery, sales, labeling, and QA move faster without engineering babysitting.
01 Product Building
• Take fuzzy problems from a conversation to a shipped v1 quickly, then iterate with real users - internal teams today, and the experience AI-lab customers touch when they receive a delivery.
• Make pragmatic build-vs-buy and scope calls; ship the 80% that unblocks people now and harden it as it proves out.
Data & Platform Plumbing
• Build the pipelines and APIs that package, validate, index, and serve ego datasets - the systems that make petabyte-scale video and sensor data usable for delivery, labeling, QA, and model training.
• Keep the tools you ship reliable and honest about data quality, so problems get caught before a dataset reaches a customer.
Who You Are
• 3+ years building software as a full-stack or product engineer (or equivalent impact) - you're comfortable owning a feature from UI to data.
• Fluent in a modern front-end stack (e.g., TypeScript/React) and at least one backend language (e.g., Python, Node, Go).
• A demonstrated 01 builder - you've taken tools or products from nothing to in-use, and you thrive in ambiguity without a detailed spec.
• Strong product sense and a bias to ship: you scope ruthlessly, get a v1 in front of users, and iterate.
• Comfortable working directly with data - querying databases (e.g., SQL, MongoDB) and reshaping or validating datasets.
Even better if you have:
• You've been an early or founding engineer, built internal tools teams loved, or shipped side projects end to end.
• You have real UX/design instincts - your tools are fast, clear, and pleasant to use, not just functional.
• You think in terms of the user, clean interfaces, and shipping - and you write maintainable code with strong ownership.
• Exposure to video/media or large-dataset systems (ingestion, storage, metadata/search, ETL) is a plus, not a requirement.
Why This Role
• Own real surface area from day one - build the products that power petabyte-scale robotics datasets and the deliveries frontier AI labs depend on.
• Maximum breadth: front-end, back-end, data, and product, on problems that are still greenfield.
• High ownership and direct impact across delivery, operations, sales, and research - paired tightly with the TPM who owns the delivery programs you build for.
A Note on Applying
Studies 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.
