The company is building systems that learn for themselves rather than relying on hand-coded expert knowledge, aiming to automate the scientific method end to end across medicine, materials, and energy. It combines advanced AI models with its own AI Science Factory instruments.
The Open-Endedness team, led by Ken Stanley, studies how a continuous chain of deeply transformative creativity can be sustained well beyond what today's models produce by imitation. The goal is systems that move past solving user-posed problems and start proposing new directions for science itself.
The team is looking for a wide range of machine learning expertise to support bold, unconventional research. Areas named include pre-training, fine-tuning, RLHF, distillation, mechanistic interpretability, Bayesian and statistical methods, and quality diversity (QD) techniques, though this list is not exhaustive.
The company operates with startup speed and holds itself to the values of truth, trust, curiosity, grit, and velocity. It encourages applications even from candidates who do not match every qualification.
Responsibilities
Design, build, and change generative models such as LLMs, diffusion models, and multimodal models using non-standard pipelines aimed at non-standard behaviour.
Develop evaluation methods that go beyond the usual, including subjective judgement and assessing how interesting an output is.
Find new ways to probe, interpret, and visualise the internal representations of large models. This covers mechanistic interpretability and approaches that go past it.
Work on quality diversity (QD) algorithms such as MAP-Elites, novelty search with local competition, POET, OMNI, and minimal criterion novelty search, applied to LLMs or other large models, which may be updated within the inner loop.
Requirements
Substantial academic or industry experience in machine learning. A PhD in a quantitative field is preferred, but self-taught researchers with exceptional achievements will be considered.
Publications at venues such as NeurIPS, ICML, AAAI, ICLR, GECCO, and ICCC.
Working knowledge of ML frameworks (PyTorch, TensorFlow, or Jax), and/or experience implementing QD algorithms, and/or neuroevolution algorithms.
World-class depth in at least one area of machine learning.
Experience training and deploying models on distributed computing platforms such as AWS, GCP, Azure, or compute clusters.
Strong written and verbal communication.
Conditions
Full-time, based in the United States or abroad. Outside the U.S., pay is set to the local market and USD ranges apply only to U.S.-based roles.
Expected base salary range: $288,000 to $480,000 USD.
Final offer depends on your background, expertise, and expected impact.
Equal employment opportunity is offered regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender identity, or Veteran status.
Benefits
Competitive base pay, bonus potential, and early-stage equity.
U.S. full-time staff receive medical, dental, and vision cover, employer-paid life and disability insurance, flexible time off with generous company-wide holidays, paid parental leave, an education assistance programme, commuter benefits including bike share memberships for office-based staff, and a subsidised lunch programme.
Full-time staff outside the U.S. receive a benefits package tailored to their region.