AI Resident

Odyssey • Zurich, California

10d • Onsite

About The Position

The AI Residency is built for a different kind of person: someone who's likely post-MS or equivalent, with real evidence they can do serious ML work, but who isn't yet operating at the level of independent, publication-driving science. You may not have direct experience in world models, generative video, multimodal learning, robotics, or multi-agent systems specifically, what matters more is that you're sharp, relentless, and capable of building that expertise fast, embedded alongside people who already have it. Residents work directly inside our core team projects, not on a side track. The work doesn't need to be publishable. It needs to move our actual research roadmap forward, and it needs to turn you into a stronger researcher and engineer by the time you leave. Residents who excel over the course of the program are extended full-time offers to join Odyssey permanently.

Requirements

• Post-MS or equivalent experience, with some demonstrated evidence of ML research ability (coursework, a thesis, prior internships, competitive results, open-source contributions, or similar)
• Strong fundamentals in machine learning and software engineering, and genuinely fast at picking up new subfields
• Not necessarily experienced yet in world models, generative video, multimodal learning, robotics, or multi-agent systems, but hungry to build that expertise on the job
• Comfortable with ambiguity and a fast-moving, high-ownership environment; you don't need everything specified for you to make progress
• Able to work in person at one of our offices in Palo Alto, London, or Zurich for the duration of the residency
• Genuinely motivated by Odyssey's mission: discovering the GPT-3 moment for world models

Responsibilities

• Work embedded within one of our research pods on a live project, whether that's world simulation, multimodal generation, robotics, or multi-agent systems
• Take on real, scoped pieces of our roadmap under the guidance of senior researchers and engineers, not isolated side projects
• Run experiments, build and debug training and evaluation pipelines, and iterate quickly on ideas with real compute behind them
• Contribute to the models we ship, and see your work show up in systems used by real people
• Learn the full loop of frontier research: hypothesis, implementation, experimentation, and honest evaluation of what worked and what didn't

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level