Data Scientist for Scientific AI Evaluations

Periodic Labs; Montreal (administrative region); On-site; CAD 347,000 - 486,000; Full time

Job summary

Periodic Labs in Montreal, Canada, is seeking a scientist to own evaluation and data across the training stack, building datasets, environments, and pipelines for cutting-edge AI models.

You will collaborate with computational and experimental scientists to translate workflows into rigorous evals and benchmarks, delivering a tight feedback loop between use cases, model evaluation, and training data.

Qualifications

• Designed evaluations, benchmarks, or RL environments for language models, agents, or scientific AI systems.
• Built large-scale data pipelines for LLM pretraining, midtraining, post-training, or evaluation.
• Strong judgment about dataset quality, provenance, licensing, and contamination risks.
• Strong software and data engineering skills, including data processing at scale and dataset versioning.

Responsibilities

• Own the evaluation and data strategy across the training stack, identifying capability gaps and shaping the roadmap with leads of physical science and AI research.
• Translate advanced scientific workflows into rigorous evals, benchmarks, and RL environments.
• Source, evaluate, and procure external datasets across chemistry, physics, materials science, mathematics, simulations, and laboratory instrumentation.
• Build robust pipelines to ingest, clean, and transform for training large-scale datasets from heterogeneous sources.
• Create tooling and workflows for researchers to inspect data and diagnose failures.

Skills

Evaluation design

Data engineering

RL environments

Dataset versioning

Scientific AI

Software engineering

Education

Bachelor's degree

Job description

Periodic Labs in Montreal, Canada, is seeking a scientist to own evaluation and data across the training stack, building datasets, environments, and pipelines for cutting-edge AI models.

You will collaborate with computational and experimental scientists to translate workflows into rigorous evals and benchmarks, delivering a tight feedback loop between use cases, model evaluation, and training data.