Imaging AI Scientist

Roche•South San Francisco, CA

4d•$124,800 - $231,800•Onsite

About The Position

The new Computational Sciences Center of Excellence (CoE) is a strategic, unified group whose goal is to harness the transformative power of data and Artificial Intelligence (AI) to assist our scientists in both pRED and gRED to deliver more innovative and transformative medicines for patients worldwide. Imaging Data Insights, within the Computational Biology and Medicine (CBM) department, turns imaging data into decisions that move drug development forward. We develop clinical imaging endpoints and advance imaging algorithms from early exploration through validation, helping Genentech/Roche bring better medicines to patients faster. As the South San Francisco imaging hub for teams across Genentech/Roche, we work where imaging problems become tools and insights that shape real scientific and clinical decisions. We're seeking an Imaging AI Scientist to join Imaging Data Insights, applying machine learning and modern AI to turn imaging data into reusable tools, scalable workflows, and better decisions across drug development. You'll move imaging analysis beyond individual projects into reusable pipelines that support portfolio decisions, modernizing traditional and ML-based approaches and increasingly exposing them as agent-callable tools on cloud infrastructure that scales across modalities. Some of these methods may be used in clinical trials or advance under GxP regulations. Working across discovery, preclinical, and clinical teams, you'll help develop imaging solutions from endpoint selection through analysis, and partner with engineering, product, and scientific teams so the methods you build get adopted across programs. As a subject-matter expert, you'll apply ML to high-value problems across disease areas and stages of development, with room to advance new imaging models where it where it matters most. This role will be open to two levels. We're equally interested in someone ready to operate at the Senior level now and someone earlier in their career with a clear trajectory toward it.

Requirements

Demonstrated excellence with current AI/ML technology, turning ambiguous problems into effective, modern solutions that work in practice.
Experience translating analytical or ML methods into reusable code, tools, or pipelines adopted by others beyond the original author.
Track record delivering models or analyses that perform in real-world settings, with strong intuition for data quality, labeling, evaluation, and reproducibility.
Proficiency in Python and modern ML frameworks (e.g., PyTorch), with clean, maintainable, reusable code.
Demonstrated ability to ramp into and operate with discipline in a rigorous, high-stakes domain.
Ability to work at the interface of imaging science, ML, and engineering, and drive work to real-world impact.

Nice To Haves

Hands-on experience with clinical imaging data (e.g., MRI, CT, PET, OCT) or tissue-based imaging (e.g., digital pathology, spatial transcriptomics, spatial proteomics).
Experience developing quantitative imaging strategies for clinical studies, such as endpoint selection, analysis plans, and data quality.
Experience developing or integrating AI-enabled automation, agentic workflows, or decision-support tools for scientific analysis.
Experience building ML or computational workflows, including data pipelines, evaluation frameworks, deployment patterns, monitoring, or MLOps practices.
Experience with advanced ML such as multimodal modeling, representation learning, generative modeling, or interpretability, particularly in applied or regulated settings.
Familiarity with regulatory or validation considerations for clinical applications, such as biomarker validation, fit-for-purpose evaluation, or GxP practices.

Responsibilities

Build reusable tools, pipelines, and agent-callable workflows that scale imaging methods and multiply what partner teams can deliver.
Develop, evaluate, and apply ML methods that produce reproducible, interpretable, and/or actionable imaging-derived insights.
Translate biological, translational, and clinical questions into fit-for-purpose imaging and analysis strategies.
Advance new imaging models and methods where they strengthen the science.
Partner across biology, translational, clinical, and computational teams to understand the questions behind therapeutic programs and put solutions into practice.
Contribute to multimodal approaches that integrate imaging with biological, clinical, or translational evidence.