Interpretable AI Biotech Scientist — Glass-Box Models

Output Biosciences; San Francisco (CA); On-site; USD 180,000 - 280,000; Full time

Benefits offered by this job

Competitive salary and equity

Medical, dental, and vision coverage

Job summary

Output Biosciences is building a biological reasoning model that understands biology at scale and enables new therapies. You will continue developing methods to understand what the foundation model learns about biology and build tools to make its reasoning a glass box.

You will probe and reverse-engineer representations, design experiments to map capabilities, and create production-ready interpretability tools for researchers and downstream systems.

Qualifications

• PhD in CS/ML/related field with 2+ years post-doc or industry research experience, or a Bachelor's/Master's with 5+ years hands-on research/engineering in interpretability.
• Strong publication record at top venues (NeurIPS/ICML/ICLR) in mechanistic interpretability or related areas.
• Hands-on experience analyzing internal representations of large neural networks.
• Proficient in Python and PyTorch; experience with large models on GPU infrastructure.
• Production-quality code experience and working in shared codebases with version control.
• Ability to design experiments distinguishing real capabilities from artifacts.

Responsibilities

• Develop methods to probe and reverse-engineer the model's learned representations across molecular scales.
• Design and run experiments to map what the model has learned about molecular interactions and biological function.
• Build tools to extract the model's biological understanding as explicit outputs for researchers and downstream systems.
• Create tools connecting model internals to meaningful biological concepts for scientists and practical use.
• Collaborate with pretraining and generation teams to feed interpretability findings back into model development.
• Own the full pipeline from probing experiments to production-quality interpretability tools on distributed infrastructure.

Skills

Python

PyTorch

Model interpretability

Representation analysis

Experiment design

Production-quality code

Education

PhD in CS/ML/Physics/Math

Bachelor's or Master's with 5+ years in research/engineering

Tools

Git

Job description

Output Biosciences is building a biological reasoning model that understands biology at scale and enables new therapies. You will continue developing methods to understand what the foundation model learns about biology and build tools to make its reasoning a glass box.

You will probe and reverse-engineer representations, design experiments to map capabilities, and create production-ready interpretability tools for researchers and downstream systems.