Lead Generative ML Scientist, Molecular Design - Equity

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

Benefits offered by this job

Competitive salary and equity

Excellent medical, dental, vision

Job summary

Output Biosciences is seeking a leader to design and develop generative models across molecular modalities, from research to deployed systems. You will architect architectures for small molecules, peptides, and beyond, and run distributed multi-GPU training.

You will own end-to-end training, evaluate biology-relevant metrics, and ship production-quality code. A PhD or equivalent with 2+ years postdoc or 5+ years experience in generative modeling is expected.

Qualifications

• PhD in CS/ML/related field with 2+ years post-doc or industry research
• 5+ years hands-on research in generative modeling (Bachelor/Master route)
• Strong publications in top venues (NeurIPS/ICML/ICLR)

Responsibilities

• Design and build generative architectures for molecular data spanning multiple modalities
• Lead end-to-end training on distributed multi-GPU clusters
• Develop controllable generation methods with biological constraints
• Evaluate generated molecules for biological relevance and novelty
• Ship production-quality code in shared repositories and contribute to open-source
• Own experimentation lifecycle from design to deployment

Skills

Python

PyTorch

Distributed GPUs

Generative models

Research experience

Publication track record

Education

PhD in CS/ML/Physics/Maths

Bachelor's/Master's with 5+ years hands-on research

Tools

Git

Docker

Job description

Output Biosciences is seeking a leader to design and develop generative models across molecular modalities, from research to deployed systems. You will architect architectures for small molecules, peptides, and beyond, and run distributed multi-GPU training.

You will own end-to-end training, evaluate biology-relevant metrics, and ship production-quality code. A PhD or equivalent with 2+ years postdoc or 5+ years experience in generative modeling is expected.