Staff AI Scientist, Foundation Models

Tempus AI•New York City, NY

About The Position

Tempus is seeking a Staff AI Scientist, Foundation Models to join their Foundation Model group. This multidisciplinary team focuses on developing foundation models across diverse clinical and molecular modalities, including longitudinal patient data, pathology, genomics, and other biomedical data. The goal is to build generalizable AI representations and models that can power applications across clinical care, biomedical research, and the drug development lifecycle. As a Staff AI Scientist, you will be a senior individual contributor focused on advancing large-scale foundation models for precision medicine, working at the intersection of machine learning, clinical science, and life sciences. You will develop novel models using one of the world’s largest clinical and molecular datasets and translate them into impactful applications across clinical research, drug development, and partner programs.

Requirements

Ph.D. in Computer Science, or a related quantitative field.
3+ years of industry experience driving end-end development of machine learning models, from research and experimentation to validation and production deployment.
Demonstrated expertise in large-scale deep learning and foundation model development is required, including experience in areas such as pretraining, representation learning, self-supervised learning, multimodal modeling, or post-training.
Proficiency in Python and the modern deep learning stack (e.g. Pytorch).
Strong product-level coding skills for model development are essential, along with a solid understanding of industry-standard software engineering and development practices.
A strong track record of original, high-impact research demonstrated through peer-reviewed publications in leading machine learning, computer vision, NLP, or related AI venues is required.

Nice To Haves

Experience with computer vision, NLP, pathology, longitudinal EHR, genomics, or other biomedical modalities is preferred.
Prior work in medical or healthcare AI is highly valued but not required.
Publications in medical or healthcare AI and prior experience developing foundation models are strongly preferred.

Responsibilities

Serve as a senior individual contributor within the foundation model team, providing strong hands-on technical contributions while helping shape model architecture, research direction, and execution.
Work closely with scientists and engineers to take ambitious AI technologies from research through validation and deployment.
Design, train, and advance state-of-the-art foundation models across large-scale multimodal clinical and molecular data, including longitudinal patient records, pathology, genomics, and related biomedical modalities.
Drive innovation in areas such as pretraining, representation learning, self-supervised learning, multimodal learning, model scaling, and post-training.
Develop and adapt foundation-model-derived models for downstream clinical and life sciences applications, including prognosis, treatment outcome prediction, patient stratification, biomarker discovery, clinical trial applications, and partner-specific AI solutions.
Contribute original ideas that advance Tempus’s foundation model research agenda, publish impactful research, and engage with the broader AI research community.
Work with cross-functional teams to ensure models and AI products meet appropriate quality, validation, and regulatory standards for healthcare and life sciences applications.