Member of Technical Staff, Pretraining Science
Core
Develop new pretraining methods, study scaling behavior, and design training recipes for large-scale biological world models.
Role type
Senior IC pretraining science researcher (biological foundation models)
Builds
Biological foundation models (multimodal genome language models)
Domain
AI research + Computational Biology
Deliverable
production ML models
Required skills
Large-scale model training, pretraining, representation learning, optimization, scaling laws, Python, PyTorch, distributed computing, experimental design, data curriculum design
Preferred skills
Frontier/foundation model training, probability/statistics/optimization fundamentals, curriculum learning, evaluation framework design, open-source contributions, applied math/systems/computational biology background
Technologies
PyTorch, distributed computing environments
Responsibilities
Research and develop new pretraining methodologies for multi-modal biological data; Study scaling behavior regarding model size, data composition, and compute budget; Design data curricula and sampling strategies; Work on architecture, algorithms, and optimization for training stability; Run large-scale experiments with empirical rigor; Collaborate with infrastructure and data teams; Define evaluations for pretraining progress