Machine Learning Researcher - Models

The Bot Company

We're building a helpful robot for every home.

We're a small team of engineers, designers, and operators based in San Francisco. Our team comes from Tesla, Cruise, OpenAI, Google, Pixar, and many other great companies. In the past we've shipped to hundreds of millions of users and know what it takes to build amazing products and experiences.

Our team is deliberately lean to promote rapid decision making and do away with bureaucracy and hierarchy. Everyone is an IC and is empowered with massive scope, radical ownership, and direct responsibility. We work across the stack with a culture built for rapid iteration and fast execution.

All roles at The Bot Company demand extreme sharpness and the ability to move fast in high‑intensity environments. Throughout the process, we expect candidates to demonstrate:

• Exceptional mental acuity: you think quickly, learn instantly, and reason across unfamiliar domains.
• Engineering curiosity: you naturally dig into how systems work, even outside your specialty.
• High performance mindset: you move fast, handle ambiguity, and excel when the environment is demanding.

We are building neural simulators that understand the grammar of the physical world—including physics, causality, and long‑term dynamics.

This role focuses on pushing video generation systems beyond short clips into controllable, large‑scale world models that can simulate environments, actions, and interactions over long time horizons. These models will serve as the foundation for robotic intelligence, enabling robots to reason about the future, anticipate consequences, and learn from simulated experience.

You will work on the frontier of generative modeling and large‑scale training to build spatiotemporal models capable of learning coherent world dynamics.

• Architect Neural Simulators:
  Design and train large‑scale spatiotemporal models capable of learning long‑horizon dynamics and physical interactions.
• Build World Models:
  Develop controllable video generation systems that evolve from short clips into coherent, persistent simulations of the real world.
• Scale Training:
  Train and optimize multi‑billion parameter models across massive GPU clusters.
• Own the Training Loop End-to-End:
  Design, run, debug, and iterate on large‑scale training experiments—diagnosing failure modes, improving data mixtures, and refining evaluation.
• Push Model Architecture Forward:
  Develop novel approaches for scaling temporal coherence, memory, and cont rollability in generative models.
• Work Across the Stack:
  Collaborate with infrastructure, robotics, and autonomy teams to integrate world models into broader robotic intelligence systems.

• Very strong coding skills in Python, C++, or Rust.
• Video Generation Expertise: deep experience building or researching high‑fidelity video generation systems.
• Architectural Intuition: ability to design novel model architectures and reason about scaling laws, emergent behavior, and failure modes.
• Infrastructure Fluency: comfortable managing large‑scale experiments across massive GPU clusters.
• Strong understanding of modern generative modeling techniques (diffusion, transformers, autoregressive models, or related approaches).

You’ll work with a small, elite team on challenges that require speed, intelligence, and deep engineering instinct. If you enjoy understanding systems at all levels, move fast, and think even faster, you’ll thrive here.

Location:

San Francisco

Employment Type:

Full time Location Type:
On‑site Department:
Engineering / Software

Compensation:
Base $200K – $350K Actual compensation will depend on skills, experience, and qualifications.

Base salary is one part of the total compensation package. The role is also eligible for equity through the company’s discretionary equity program, along with a comprehensive benefits package that includes medical, dental, and vision coverage, and access to a 401(k) plan.