Member of Technical Staff, AI

Pathos•New York, NY
•$220,000 - $350,000

About The Position

The Member of Technical Staff, AI will advance the foundation model and build agentic systems behind Foundry, Pathos’s AI platform for informing drug-development decisions. The role focuses on adapting frontier open-weight models for drug development and improving them over time using real-world feedback. The work draws on more than 200 petabytes of multimodal, multi-scale oncology data spanning over 8 million patient records linked to longitudinal outcomes, as well as live data from clinical trial sites. The position is supported by a large-scale GPU cluster and an internal wet lab for experimental validation, and its models will directly inform the design and execution of clinical trials. This is an end-to-end ownership role on a lean, high-velocity team. The successful candidate will work with oncologists, biologists, and clinical colleagues to translate drug-development needs into tractable machine-learning problems, and will communicate and publish results in leading ML and scientific venues. Pathos is hiring across experience levels, from exceptional early-career talent to experienced researchers. The preferred location is the New York City headquarters, though exceptional remote candidates will be considered. Applications are reviewed continuously, with priority consideration for submissions received by December 7, 2026. The anticipated salary range is $220,000 to $350,000 plus equity, with flexibility for exceptional candidates.

Requirements

  • Deep, hands-on expertise in at least one of the following areas: mid- or post-training at scale (such as SFT, RL, reward modeling, or preference optimization) with direct ownership of training runs and results; or evaluation, data, and agent-harness design for scientific, clinical, or operational tasks.
  • Rigorous scientific judgment, including when ground truth is limited, delayed, or disputed.
  • Strong Python engineering skills and the ability to write production-quality code.
  • Clear communication of evidence, limitations, and uncertainty.
  • Willingness to learn biology and drug development and work closely with domain experts.

Nice To Haves

  • Experience building self-improving agentic systems, such as through self-play, synthetic data generation, self-training on past trajectories, test-time learning, or systems that automate ML research.
  • Experience building AI systems for real-world scientific workflows, such as agents that generate hypotheses and design, run, and analyze experiments using laboratory or clinical data.
  • Experience building RL environments or task suites for training agents on realistic, long-horizon tasks.
  • Experience modeling multimodal biomedical data, such as genomics, transcriptomics, pathology and radiology images, clinical notes, and EHRs, ideally combining several modalities in one model.
  • Experience with uncertainty estimation, calibration, or causal inference, especially for high-stakes decisions.
  • Experience building or scaling infrastructure for efficient training and inference, such as distributed training, custom CUDA or Triton kernels, or high-throughput serving.
  • Experience shipping ML systems used by real users, monitoring them in production, and improving them using user feedback.
  • Background in oncology, drug development, or clinical research, or a track record of working closely with clinicians and biologists.
  • Experience with patient-level or clinical trial data under privacy and regulatory constraints, such as HIPAA.
  • Publications, research software, or deployed systems as evidence of relevant work; publications are welcome but not required.

Responsibilities

  • Advance mid- and post-training of the oncology foundation model across supervised fine-tuning and reinforcement learning, including reward design and verifiers for scientific and clinical tasks that are difficult to check automatically.
  • Run and debug large-scale training, addressing instability, divergence, and reward hacking.
  • Build the harness for scientific and clinical workflows, including memory, context management, orchestration, and tools, and develop methods to optimize the harness and evolve it with the model.
  • Design evaluations for open-ended, long-horizon scientific, clinical, and operational tasks; calibrate them against expert judgment, investigate failures, and use them to inform release decisions.
  • Build pipelines to generate, filter, and curate training and evaluation data, with safeguards for quality, provenance, and contamination.
  • Turn expert feedback, wet-lab results, and real-time signals from clinical trial sites into improvements to models, the harness, and evaluations.
  • Work with oncologists, biologists, and clinical colleagues to translate drug-development needs into tractable ML problems.
  • Publish in leading machine-learning and scientific venues.

Benefits

  • Equity
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service