We hire deeply technical staff working in world models, video generation, multimodal, robotics, autonomous vehicles, and adjacent fields. Applied Research sits where the models meet real applications. Whether you're building from scratch, scaling training runs, or working on inference, we hire people who can take an idea and make it real.

You'll lead our ML platform and infrastructure group, spanning our data platform, evaluation systems and tooling, and compute. This leader will serve as a trusted partner to the founders and stakeholders across product and technology, and will report directly to our CTO and co-founder.

Data is fast becoming one of the biggest bottlenecks in building world models. Our models are only as good as the data behind them, and getting that data right is one of the hardest and most important problems we have. We're looking for a data engineer who wants to be the person figuring it out: building and curating the large-scale multimodal datasets our models train on, across video, robotics, and audio. The work spans research and infrastructure. Some days you'll be tuning the platform that processes millions of hours of video and audio. Other days you'll be in the data itself: pulling out signals, improving captions, and working with researchers on what goes into the training mix. We don't treat those as separate jobs, so we want someone comfortable doing both. We care more about how you think about data than about your years of experience or publication record. You might have started in data engineering and moved toward research, or started in research or data science and moved toward engineering. Either way, you like working on data and you've done hands-on work with video and/or audio.

The right person will have a deep interest in improving machine learning algorithms. Real-time world model simulation is a cutting edge research area that is not yet mature. You will be working at the cusp of what's possible. Most new experiments in this area will fail, your focus will be on maximally learning from failed experiments to increase the chances of eventual success.