Scientist, Predictive Biology and AI

Role Summary

Scientist, Predictive Biology and AI. Location: Seattle, Brisbane CA, San Diego CA, Cambridge MA, Princeton NJ. The Predictive Biology and AI (PBAI) team develops and applies cutting-edge methods to address patient needs in Oncology, Neuroscience, and other areas. The role focuses on evaluating and adapting state-of-the-art AI models to challenges in cell engineering and target discovery, collaborating with wet-lab partners to test predictions and integrate data into models.

Responsibilities

• Apply, adapt, and in some cases create multi-modal foundation models such as large language models (LLMs), diffusion models, and encoder architectures to answer biological domain-specific questions
• Address real-world biological modeling challenges such as data sparsity, class imbalance, noise, experimental bias, and heterogeneity of effects
• Thoughtful model evaluation that incorporates appropriate benchmarks, statistical tests, and problem understanding to support technical and business decisions
• Work in close collaboration with partners across the organization including wet-lab scientists, Research IT, and other computational scientists to broaden the impact of AI developments
• Maintain and share up-to-date knowledge of modern advances in the field, including presenting work at public conferences

Qualifications

• Required: Bachelor's Degree with 5+ years of academic/industry experience; or Master's Degree with 3+ years; or PhD (no experience required)
• Preferred: Ph.D. with 0+ years or M.S. with 3+ years in computer science, statistics, computational biology, or another quantitative field
• Expert-level understanding of deep learning tools and approaches (transformer encoders/decoders, LLMs, reinforcement learning) demonstrated through publications or projects
• Hands-on experience building and scaling deep learning training pipelines on multi-GPU infrastructure using PyTorch, Huggingface, and related tools
• Knowledge of or ability to learn biological concepts and data types, with ability to work and communicate effectively with biologists
• Excellent verbal and written communication skills in English
• Experience building agentic workflows is a plus; prior experience in pharmaceutical applications is a plus

Skills

• Machine learning and statistics
• Multi-modal AI models (LLMs, diffusion models, encoders/decoders)
• Data analysis with attention to sparsity, bias, and heterogeneity
• Collaborative teamwork with wet-lab and computational partners
• Communication of complex AI concepts to non-technical stakeholders