ML Research Scientist, Scientific Evaluations

Lila Sciences•Cambridge, MA
•$176,000 - $304,000

About The Position

Lila Sciences is hiring an ML Research Scientist for its AI Research team to build scientific problems and environments that evaluate frontier LLMs. The role focuses on designing and prototyping science evaluations, diagnosing environment failures, and addressing model capability gaps. It also involves inventing RL environments that convert scientific workflows, simulators, and real-world tasks into training signals. The scientist will build end-to-end pipelines that transform raw assets such as research papers, codebases, and lab protocols into validated, versioned training environments at scale. They will prototype and validate environment designs from initial concept through reward shaping and data collection, and integrate them into the training loop. The position applies AI and automation to scientific discovery across fields including medicine, materials, and energy.

Requirements

  • Experience applying machine learning or computation to biology, chemistry, or material sciences.
  • A strong track record of creating environments or evaluations for model training.
  • Experience building large, complex data pipelines for model training or production.
  • Strong prototyping and coding skills, with the ability to quickly build and iterate on ideas in Python.
  • Creativity and research intuition, including the ability to identify promising directions, design experiments, and navigate ambiguity.
  • Solid software engineering fundamentals, including clean code, version control, testing, and integrating experimental work into shared infrastructure.

Nice To Haves

  • Demonstrated experience training large language models.
  • A strong publication track record at conferences such as NeurIPS, ICML, and ICLR.
  • Strong Python programming skills, including research and production infrastructure.
  • Familiarity with distributed compute systems such as Ray or Modal, or LLM inference systems such as vLLM and batching strategies.
  • Experience with async Python, Pydantic, and workflow orchestration frameworks such as Flyte or Airflow.

Responsibilities

  • Design and prototype evaluations and environments for science, including identifying environment failures and addressing model capability gaps.
  • Invent RL environments that turn scientific workflows, simulators, and real-world tasks into training signals.
  • Build end-to-end pipelines for generating synthetic environments from papers, codebases, lab protocols, and other raw assets.
  • Prototype and validate environment designs from initial concept through reward shaping and data collection.
  • Integrate environment designs into the training loop.

Benefits

  • Competitive base compensation with bonus potential and generous early-stage equity.
  • Medical, dental, and vision coverage for full-time U.S. employees.
  • Employer-paid life and disability insurance for full-time U.S. employees.
  • Flexible time off and generous company-wide holidays.
  • Paid parental leave.
  • Educational assistance program.
  • Commuter benefits, including bike share memberships for office-based employees.
  • Company-subsidized lunch program.
  • Region-tailored benefits program for full-time employees outside the U.S.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service