Research Engineer - Midtraining

Periodic Labs
Menlo Park, San Francisco
Workplace: OnsiteFull timeUSD 250,000 - 350,000 annuallyFunction: Research & Scientific (R&D)Education: bachelorsSkills: ["Ownership","Drive","Collaboration"]

Train frontier AI models to improve scientific reasoning for discovery. Curate and generate novel and synthetic scientific data, build evaluations tied to downstream scientific performance, and run large-scale training experiments with partners in RL and domain science. Apply techniques like self-distillation and on-policy distillation, and collaborate with supercompute engineers to efficiently scale across thousands of GPUs while developing tools to analyze how data choices shape model intelligence.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Periodic Labs
Periodic Labs
1 day ago

Research Engineer - Midtraining

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 7 hours agoStatus: Live

Job Summary

Train frontier AI models to improve scientific reasoning for discovery. Curate and generate novel and synthetic scientific data, build evaluations tied to downstream scientific performance, and run large-scale training experiments with partners in RL and domain science. Apply techniques like self-distillation and on-policy distillation, and collaborate with supercompute engineers to efficiently scale across thousands of GPUs while developing tools to analyze how data choices shape model intelligence.
Location: Menlo Park, San Francisco
Workplace: Onsite
Employment Type: Full time
Job Function: Research & Scientific (R&D)
Seniority: Mid level

Key Responsibilities

  • •Identify, process, and curate novel sources of scientific data for large-scale model training.
  • •Generate high-quality synthetic data to fill gaps in scientific knowledge and reasoning.
  • •Build evaluations that correlate with downstream scientific task performance, partnering with RL researchers and domain scientists.
  • •Develop and apply techniques such as self-distillation and on-policy distillation to improve model capability.
  • •Design and run large-scale training experiments and build tools to analyze how data choices shape model intelligence.

Pay and Benefits

Salary: USD 250,000 - 350,000 annually
Perks:Equity

Key Requirements

  • •Experience training LLMs on curated mixes of trillions of tokens.
  • •Hands-on experience with self-distillation, on-policy distillation, or similar methods in a real training pipeline.
  • •Experience on a dedicated evals team supporting a large production training run.
  • •Experience with scaling laws and compute-optimal hyperparameters.
  • •Comfort working across data, evals, and training infrastructure.
Experience:AI for scienceMachine learning
Education:Bachelor's
Skills:OwnershipDriveCollaboration
Tech Stack:LLMsSelf-distillationOn-policy distillationEvaluations (evals)Large-scale trainingGPUsDistributed trainingSynthetic dataRLTraining pipelines

Company Brief

Periodic Labs
Periodic Labs (periodic.com) — Public information about this company’s products, industry focus, size, funding, and headquarters is not available. Check the company’s official profiles or contact them directly for accurate details about their offerings and operations.
Industry: Other
Website