Member Of Technical Staff, Pretraining Science

Radical Numerics

San Francisco, US
On-site
Large-scale model training
Pretraining methodologies
Representation learning
Radical Numerics is building the infrastructure to unlock scaling on vast biological sequence, structure, and image datasets for biological world models

Job Summary

  • Radical Numerics is building the infrastructure to unlock scaling on vast biological sequence, structure, and image datasets for biological world models.
  • This role focuses on developing new pretraining methods, studying scaling behavior, and designing training recipes to improve efficiency and generalization for biological world models.
  • The company offers the opportunity to work on fundamental questions in pretraining science while staying close to real scientific applications in biology.

Matching Summary

Radical Numerics is building the infrastructure to unlock scaling on vast biological sequence, structure, and image datasets for biological world models.

Skills & Requirements

Must-have

  • large-scale model training
  • pretraining methodologies
  • representation learning
  • scaling laws
  • Python and PyTorch proficiency
  • distributed training systems

Nice-to-have

  • frontier or foundation models
  • probability and statistics
  • curriculum learning
  • open-source ML contributions
  • quantitative field background

Key Requirements

  • ML research or engineering track record
  • experimental design and analysis skills
  • Python and deep learning tooling experience
  • distributed or HPC environments experience
  • strong communication skills

Work Rights

Not specified

Tailored Resume

Cover Letter