Model Behavior, Model Designer

The mission of Thinking Machines is to build AI that extends human will and judgment. We are training frontier models with Inkling, developing Tinker to let people make models their own, and crafting interfaces that broaden human-AI communication. We believe the future worth building is human, and we're hiring people who want to build it.

About the Role

As a Model Designer, you'll shape how our models work alongside people. Our mission is to build AI that extends human will and judgment, and this role is where that mission meets the actual texture of an interaction: how a model reads intent, when it acts on its own, when it asks, and whether the person comes away sharper than they started.

You'll hold several things in tension: pushing what our models can do, anticipating what someone is reaching for before they've fully said it, and keeping the person's judgment at the center of the work rather than quietly replacing it. We're looking for people who care deeply about design, technology, and the humans on the other side of the screen, and who want to help invent forms of human-AI collaboration that go well beyond typing into a box and waiting.

What You’ll Do

• Work closely with researchers to understand, predict, and shape model behavior, knowing that lasting behavior lives in the weights, not just the prompt.

• Design how our models collaborate: when to take initiative, when to invite oversight or feedback, and how to earn the trust to act on someone's behalf over time.

• Build strong, thoughtful defaults that serve as a starting point people can reshape around their own knowledge and values, rather than a single voice imposed on everyone.

• Partner with research, product, and design teams on interfaces that widen the channel between people and models, including live and multimodal interaction.

• Find creative ways to collect high-quality data, especially the tacit, hard-to-articulate knowledge people build through doing real work.

• Help define what good looks like, with evaluations that capture what people and models accomplish together, not only what a model can do alone.

• Proactively identify improvements based on product sense, user feedback, quantitative insight, and the research roadmap.

• Manage multiple projects on tight timelines, and do whatever it takes to make our models better collaborators.

Skills and Qualifications

Minimum qualifications:

• Clarity in communication, an ability to explain complex technical concepts in writing.

• Demonstrated ability to understand users whose needs and contexts differ from your own, such as research across varied user populations, domains, or cultures.

• Several years of experience in a role focused on how people interact with technology, such as product design, UX research, content or conversation design, or model behavior work at an AI lab.

Preferred qualifications — we encourage you to apply even if you don’t meet all preferred qualifications, but at least some:

• Have past experience in designing model behavior or personality, or influencing technical roadmaps related to unverifiable domains.

• Past experience in crafting research taste, model values, epistemics, or philosophies

• Hands-on experience with large language models, including writing and iterating on prompts, evaluating outputs, and building an intuition for why a model behaves the way it does.

• A portfolio or writing samples that show exceptional taste and clarity, ideally including work that shaped how a product or system communicates with people.

• A track record of taking ambiguous, loosely defined problems from framing to shipped results, working across research, product, and engineering.

• Experience designing and running experiments (qualitative studies, A/B tests, or model evaluations) and making decisions based on the results, including dropping ideas the evidence didn't support.

• Comfort with quantitative analysis: you can read eval results, find patterns in data, and explain findings to both technical and non-technical audiences.

• A clear point of view on how AI should support human judgment rather than replace it, shown through your past work or writing.

Logistics

• Location: This role is based in San Francisco, California. 

• Compensation: Depending on background, skills and experience, the expected annual salary range for this position is $300,000 - $400,000 USD.

• Visa sponsorship: We sponsor visas. While we can't guarantee success for every candidate or role, if you're the right fit, we're committed to working through the visa process together.

• Benefits: Thinking Machines offers generous health, dental, and vision benefits, unlimited PTO, paid parental leave, and relocation support as needed.