Scientist /Senior Scientist, Multimodal & Relational Machine Learning Foundation Models
Core
Designing, developing, and evaluating state-of-the-art multimodal generative foundation models for multiscale biology to aid in the discovery of novel interventions for aging and disease.
Role type
Staff Machine Learning Scientist (Multimodal & Relational Foundation Models)
Builds
Unified multi-modal generative foundation models integrating LLMs with Graph Neural Networks for biological knowledge graphs
Domain
Biotechnology / Machine Learning / Graph Neural Networks
Deliverable
production ML models
Required skills
Multimodal data integration, Graph Neural Networks (GNNs), Large Language Models (LLMs), Distributed training (FSDP, DeepSpeed), Python, PyTorch/JAX, Peer-reviewed research publication
Preferred skills
Tabular foundation models, In-context learning, NGS data analysis, Biological imaging, Model optimization (quantization/distillation)
Technologies
PyTorch, JAX, Hugging Face Transformers, FSDP, DeepSpeed, Megatron, Ray, DDP
Responsibilities
Pre-train and fine-tune large-scale ML systems using multimodal biological data; Architect hybrid models integrating LLMs with GNNs; Develop Relational Foundation Models for zero-shot predictive tasks; Lead design of efficient data loading and distributed training strategies; Apply theory to model performance analysis; Transition research prototypes to production systems; Mentor junior staff
Seniority
Senior, hands-on IC with mentorship
