About the role

We're hiring an experienced ML engineer to work on the models that see. You'll own problems end to end: deciding what data you need, getting it, training on it, proving the result is actually better, and getting it running inside the vehicle's constraints.

The problems you'd be working on

Rather than a list of responsibilities, here's what the team is actually chewing on:

A model that's two points better offline can be worse on the road. Aggregate benchmark numbers hide the failures that matter — the rare scene, the unusual agent, the bad lighting. Building evaluation that predicts on-road behaviour, and knowing when to distrust your own metric, is a bigger part of this job than architecture search.

We generate far more data than anyone can look at. The interesting frames are a vanishingly small fraction of what the fleet records. Finding them, deciding what's worth labelling, and keeping the training set honest as the distribution shifts is continuous work, not a one-time setup.

The vehicle's compute budget is fixed and already full. Everything you add competes with everything already running. You'll be making concrete trades between accuracy, latency, and memory, and defending them.

Modern architectures keep changing what's possible. Transformers and multimodal models opened up approaches that weren't available two years ago. Part of the job is reading what's coming out, judging honestly whether it applies to our problem, and being willing to conclude that it doesn't.

Nothing ships alone. Your model's output is someone else's input. You'll work directly with the planning, infrastructure, and vehicle software teams, and the handoffs are where most of the real difficulty lives.

What we're looking for

• You've shipped a neural network, not just trained one. At least three years taking models from data collection through training to something that ran in production or on real hardware, and stayed working.
• Real depth in one modern ML area — computer vision, large language models, or generative modelling. We'd rather see one domain you know properly than six you've touched.
• Python and a modern deep learning framework, fluently, as your daily working environment.
• Enough C++ to be useful. Inference runs in C++ on the vehicle. You don't need to be a C++ specialist, but you need to be able to read the code your model runs inside and work with the engineers who own it.
• Comfort with large-scale data tooling and SQL — you can get your own data without waiting on someone else.
• You read papers and can tell which ones matter. Most don't.
• You can explain a technical trade-off to someone who doesn't share your background and hold your position when it's the right call.

Things that would stand out

• You've made a model meaningfully faster on target hardware and can explain what you gave up to get there.
• You've worked on ML for autonomous vehicles or robotics before, and know how different the failure modes are from a benchmark.
• Published work or open-source contributions we can actually read — send us a link and we'll read it.
• A track record of setting a direction and following it through without needing to be steered.