Research Fellowship - Open Endedness

• Develop RL methods for agents that can discover useful objectives, tasks and curricula without relying entirely on human-specified rewards.
• Design systems for open-ended learning, including unsupervised/automated environment design, asymmetric self-play, and intrinsic motivation.
• Build training loops where agents learn from interaction, exploration, novelty, competence progress, self-generated challenges, or other nonstandard reward signals.
• Investigate how agents can avoid collapse into trivial, degenerate, or easily exploitable objectives.
• Own and develop a research agenda within Vmax, from identifying promising directions to executing experiments and communicating results.

• Currently enrolled in a PhD program in machine learning, computer science, artificial intelligence, computational neuroscience, mathematics, or a related technical field. Exceptional candidates with equivalent research experience may also be considered.
• Track record of research excellence or strong research promise, demonstrated through publications, preprints, open-source work, technical projects, competitions, or publicly available artifacts.
• Working understanding of reinforcement learning.
• Familiarity with unsupervised/automated environment design, asymmetric self-play, and/or intrinsic motivation.
• Strong programming ability in Python and experience with at least one major ML framework such as PyTorch or JAX.
• Clear written and verbal communication of technical ideas.

• Experience with LLM post-training methods
• Experience with scalable ML experimentation, distributed training, experiment tracking, or reproducible research infrastructure.
• Demonstrated taste for identifying non-obvious research directions and converting them into tractable experiments.

• This role is based in our San Francisco office; for exceptional candidates we are willing to consider a hybrid arrangement

Vmax is an applied research lab developing AI capable of open-ended learning. We are building systems to exceed humans in all capacities by optimising beyond the local maxima of learning from human expertise.