About The Position

Tempus is seeking an AI Scientist/Engineer to work across oncology foundation models and agentic systems. The role involves designing and training domain-specific foundation models on complex multimodal cancer datasets, including genomics, histopathology, longitudinal EHRs, and clinical imaging, while building agentic workflows to orchestrate those models. The work aims to create intelligent clinical tools for dynamic reasoning, multi-step evidence synthesis, and complex decision support in oncology. The position also integrates oncology foundation models into agentic platforms so agents can invoke multi-omic analysis tools and support workflows such as summarizing patient trajectories, proposing biomarker-driven treatment hypotheses, and querying clinical-trial eligibility. The role emphasizes safety, interpretability, and performance for clinical and scientific research users.

Responsibilities

  • Design, train, and fine-tune large-scale multimodal foundation models specialized in oncology, including models integrating spatial transcriptomics, whole-slide images, and clinical histories.
  • Implement biomedical pre-training strategies such as self-supervised learning, contrastive learning, and masked token prediction.
  • Evaluate model robustness, generalization, and domain shift across patient cohorts, tumor types, and sequencing platforms.
  • Architect scalable agentic frameworks for clinical workflows, including tool use, multi-agent communication, autonomous planning, and retrieval-augmented generation.
  • Design safety guards, reasoning-trace verification, and human-in-the-loop interventions for agents used in high-stakes medical settings.
  • Integrate agentic systems with biological knowledge bases, literature databases such as PubMed and clinical-trial resources, and EHR interfaces.
  • Expose oncology foundation models as tools and capabilities within the agentic platform, enabling agents to invoke multi-omic analysis tools.
  • Develop agent-driven workflows for summarizing patient trajectories, proposing biomarker-driven treatment hypotheses, and querying trial eligibility.
  • Optimize end-to-end latency, reasoning chains, and output interpretability for clinical and scientific research users.

Benefits

  • Incentive compensation may be available.
  • Restricted stock units may be available.
  • Medical and other benefits may be available depending on the position.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service