Translational Scientist, Applied Machine Learning and Agentic Ai, Pharma R&d

Tempus AI Tempus AI · Vertical AI · Boston, MA +2

This role focuses on building cutting-edge agentic frameworks for automating the discovery of prognostic and predictive models in oncology, integrating LLM orchestration with computational biology. The scientist will develop 'deep agents' for hypothesis generation, experimental design, and multimodal ML modeling using foundation models, leveraging Tempus' extensive multimodal patient datasets.

What you'd actually do

  1. Develop complex, state-of-the-art agentic workflows.
  2. Build agents capable of long-horizon planning, tool use and "co-scientist" reasoning.
  3. Leverage oncology foundation models to integrate DNA, RNA, H&E, and clinical data into predictive algorithms.
  4. Collaborate with clinical scientists and pharma partners to define high-value use cases, such as clinical trial design support and treatment de-escalation.

Skills

Required

  • PhD (or Masters degree with 3+ years of relevant experience)
  • Quantitative and computational skills
  • AI agent based workflows (e.g. Applied Machine Learning, Generative AI, Mathematics, biostatistics)
  • Biological, medical, or drug development knowledge and data (e.g. oncology, RWE, medical science, or clinical drug development)
  • Proficiency in Python
  • Experience building deep agents with complex state management and graphs
  • Deep knowledge of prompt engineering
  • RAG (Retrieval-Augmented Generation)
  • Function calling
  • Evaluating non-deterministic LLM outputs
  • Strong foundation in survival analysis (CoxPH, RSF)
  • Evaluation metrics for oncology models
  • Adherence to software best practices (unit testing, git)
  • Experience designing scalable systems
  • Excellent written and verbal communication skills
  • Ability to present complex information clearly and persuasively to diverse audiences

Nice to have

  • LangGraph
  • Experience working with clinical trial or real-world data
  • Clinical guidelines (e.g., NCCN for oncology)
  • Emerging RWE methodologies
  • Integrative modeling of multi-modal clinical and omics data
  • Multimodal embeddings
  • Foundation models
  • Data and artificial intelligence in Oncology
  • Cancer biology
  • Clinical data
  • Deploying ML models in cloud environments

What the JD emphasized

  • Python and orchestration frameworks, specifically LangGraph (strongly preferred) or similar
  • Experience building deep agents with complex state management and graphs
  • Deep knowledge of prompt engineering, RAG (Retrieval-Augmented Generation), function calling, and evaluating non-deterministic LLM outputs
  • Strong foundation in survival analysis (CoxPH, RSF) and evaluation metrics for oncology models
  • Track record of success: proven in peer reviewed publications or other proven impact

Other signals

  • agentic workflows
  • foundation models
  • multimodal ML modeling
  • computational biology
  • causal inference