[CAD] Data Scientist & Engineer

Prometheus · CAD Working Group · Zurich or London or San Francisco

THE ROLE

We’re hiring a Data Scientist and Engineer to create training data for a frontier CAD and mechanical design model. You’ll work directly with researchers to turn large volumes of CAD and engineering assets into model training datasets and figure out which data actually makes the model better.

This is both a scientific and an engineering role. You’ll investigate the data, design experiments, and build the pipelines that turn your findings into better training datasets.

WHAT YOU’LL DO

- Transform raw assets into ML-ready datasets. Extraction, parsing, normalisation, deduplication, and filtering. Some transformations will be mundane, some creative; you’ll decide what’s needed.

- Understand and improve the training distribution. Analyse dataset composition, identify gaps and quality issues, and work with researchers to test how data selection, filtering, and mixing affect model performance.

- Generate and evaluate synthetic data using LLMs and VLMs. Own generation strategies, prompting, and pipelines at scale. Establish whether the resulting data improves coverage and model performance.…