Machine Learning: World Models

The role

You will develop video generation into controllable, large-scale world models. You will design and train spatiotemporal models and own the end-to-end training of multi-billion parameter models on huge clusters.

What you will do

• Architect Neural Simulators by designing and training spatiotemporal models.
• Own the end-to-end training of multi-billion parameter models on huge clusters.
• Design, run, debug, and iterate on large-scale training experiments.
• Diagnose failure modes, improve data mixtures, and tighten evaluation.

Required qualifications

• Very strong coding skills in Python, C++, or Rust.
• Deep experience shipping/researching high-fidelity video models.
• Ability to design from scratch and reason about scaling laws and failure modes.
• Comfortable managing and optimizing large-scale experiments on massive GPU clusters.

About the company

The Bot Company is a small team of engineers, designers, and operators based in San Francisco building a helpful robot for every home.