Member of Engineering (Pre-training / Data Engineering)

Stack

Responsibilities

• ▹Build and maintain high-performance pipelines for trillions of tokens
• ▹Deliver diverse and high-quality datasets for pre-training foundation models
• ▹Work closely with Pretraining, Posttraining, Evals, and Product teams to ensure model quality alignment
• ▹Engineer ingestion, deduplication, and streaming systems for petabyte-scale data
• ▹Bridge raw web crawls and GPU clusters through data modeling and distributed pipeline optimization

Requirements

• ▹Strong background in building production-grade, distributed data systems for machine learning
• ▹Experience with orchestration: Slurm, Airflow, or Dagster
• ▹Experience with observability & reliability: CI/CD, Grafana, Prometheus
• ▹Infra experience: Git, Docker, k8s, cloud managed services
• ▹Batch inference experience (e.g. vLLM)
• ▹Performance obsession at large scale

Soft skills

What we offer

• ▹Distributed team across Europe and North America
• ▹Monthly 3-day (Monday-Wednesday) in-person gathering in Paris, with an open invitation to stay the whole week
• ▹Longer offsite once a year

About the company

Poolside, founded in the US, is racing toward AGI by accelerating software development with agentic systems and coding assistants; its team is distributed across Europe and North America, meeting monthly in Paris.