Research Scientist, Pre-training Research

Chai Discovery • San Francisco, CA

2d

About The Position

Chai Discovery is hiring research scientists to work on foundation model pre-training. The models learn the foundations of molecular structure and interaction, and their success depends on carefully designed architectural components, training objectives, and data mixes. Pre-training research at Chai is largely domain-agnostic, involving understanding model architecture, reading training curves, and determining when to move between phases of a training recipe. The team favors simplification, generality, and scale, and adapts ideas from frontier language, image, and video models to the domain. Chai's models have moved beyond protein structure prediction into real-world therapeutic engineering, and the role offers compute and a team to advance this work.

Requirements

• A strong research track record, demonstrated through achievements such as getting a model to work, developing impactful theory, or making a substantial contribution to the field.
• Depth in generative modeling, model architecture, or scaling laws.
• Hands-on experience designing and training large models.
• Theoretical grounding in computer science, machine learning, mathematics, or physics.

Responsibilities

• Work on foundation model pre-training.
• Design and evaluate architectural components, training objectives, and data mixes.
• Understand model architecture and training curves, and determine when to move between phases of a training recipe.
• Adapt ideas from frontier language, image, and video models to molecular modeling and therapeutic engineering.

Benefits

• Compensation commensurate with the team’s work and contributions.
• Opportunity to work at the vanguard of AI research and frontier biology with world-class colleagues.
• A culture that promotes high velocity and ownership.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: No Education Listed