You've trained image or video generation models and shipped them to real users. Now own the models behind a fast-growing creative product.
A young AI lab in San Francisco makes video creation as easy as AI made coding, with a desktop app that brings a large library of AI models, search, motion graphics and editing into the AI assistant people already use. The product has strong traction, and the team is now building its own compact models to make it quicker and more capable. As their ML engineer, those models are yours: how they're trained, how they're tested, and which data makes them better.
What you'll own
Training and testing every model that runs in the product
Benchmarks and feedback cycles for the video models they use
Converting real user behaviour and creative taste into high-quality training and test data
Compact in-house models that speed the product up
Must-haves
You've personally trained or fine-tuned diffusion, flow-matching or similar generative models for images or video (using them via an API doesn't count)
You've taken models all the way into production. This is a builder role, not one for pure researchers or evaluation specialists.
A strong track record on the video, vision or generative-media side at a major tech company or a leading generative-media lab
A strong feel for which data actually improves a model, and how to gather and clean it
On-site in San Francisco every weekday; happy to help you relocate
Roughly 3–6 years of experience
Bonus points
A specific image or video model or feature with your name on it
Benchmarks, evaluation tooling or data pipelines you built for generative models
Publications or open source, on top of what you shipped
What's in it for you
$200k–$350k base + 0.1–0.7% equity
Full ownership of the proprietary models on a team of fewer than ten
Visa sponsorship available
Good to know
Full-time, on-site in San Francisco, 9-9-5
Process: first call with the CTO a day working together on site culture conversation with the CEO references offer
When you apply, please include
Your CV (PDF), your LinkedIn URL, and GitHub, papers or a demo if you have them
The model or feature you trained and shipped, and your own part in it
Your current location, and confirmation that on-site San Francisco works for you
Your visa or citizenship status, and if you'll need sponsorship
Your base salary expectation, your notice period, and any other interview processes you're in
Does this sound like you? Apply now. If you're a match, we'll be in touch within 3 working days.