Machine Learning Scientist, Scientific Reasoning Models, AI for Drug Discovery

About the role

A healthier future drives us to innovate, continuously advance science, and ensure everyone has access to the healthcare they need. Roche’s Research and Early Development organisations (gRED and pRED) leverage AI, data, and computational sciences to accelerate drug discovery and development. The Computational Sciences Center of Excellence (CoE) is a strategic, unified group harnessing the power of data and Artificial Intelligence (AI) to assist scientists in delivering innovative medicines.

The AI for Drug Discovery (AIDD) group at Roche is revolutionizing drug discovery with cutting-edge machine learning (ML). As a Machine Learning Scientist on the Foundation Models team within Prescient Design (gRED), you will contribute to internal reasoning Large Language Models (LLMs) and enable their success in drug discovery tasks, including biomolecular design. You will work at the intersection of engineering and research, designing and scaling large ML systems.

Responsibilities

• Scalable Systems & Engineering: Design, implement, and improve large-scale distributed machine learning systems, writing robust, performance-critical code and contributing to core infrastructure.
• Model Improvement & Reasoning: Develop and execute strategies to systematically improve performance on scientific tasks, including long-horizon task completion and complex reasoning challenges.
• Domain Translation: Translate biological and chemical domain knowledge into concrete machine learning objectives, training signals, and evaluation criteria.
• Evaluation & Benchmarks: Design and implement evaluation methodologies to assess model capabilities relevant to biological research, working with domain experts to establish benchmarks and curate high-quality data.
• Research-to-Production: Collaborate closely with researchers to translate ideas and prototypes into scalable, production-ready systems.
• Write clean, efficient code to test specific hypotheses regarding reasoning and alignment.
• Contribute to the maintenance of training infrastructure and data pipelines, ensuring experiments run reliably on clusters.
• Work closely with senior scientists to implement novel algorithms, translating research papers into working prototypes.

Requirements

• BS/MS in Computer Science, Statistics, Mathematics, Physics, or a related quantitative field with 2+ years of relevant work experience. Or Ph.D. with 0-2 years of relevant work experience.
• LLM Expertise: Experience developing and training large-scale machine learning models, including post-training techniques to enhance domain knowledge, reasoning capabilities, and model alignment.
• Publication Record: A strong history of research excellence at top-tier venues (e.g., NeurIPS, ICLR, ICML).
• Engineering: Strong software engineering skills and experience working with high-performance computing systems.

Preferred Qualifications

• Experience with molecular modalities (e.g., protein sequences, chemical graphs, and structured molecular data).
• A public portfolio of research or significant contributions to open-source ML libraries.
• A passion for applying frontier AI to drug discovery.

Pay

The expected salary range for this position is:

• $141,100 – $262,100 (New York City)
• $147,600 – $274,000 (San Francisco)

Actual pay will be determined based on experience, qualifications, geographic location, and other job-related factors permitted by law. A discretionary annual bonus may be available based on individual and Company performance.

Benefits

This position qualifies for the benefits detailed at the provided link. Relocation benefits are not available for this role.