Data / Machine Learning Engineer, Fusion Data

About the role

Fusionality, spun out of EPFL's Swiss Plasma Center and founded by experts at the intersection of fusion physics, advanced control and AI, builds the measurement, simulation and control stack for future fusion energy reactors. It is assembling a founding engineering team in Lausanne and is looking for a Data / Machine Learning Engineer who enjoys turning large, messy, heterogeneous experimental data into clean, labelled, trustworthy datasets. Familiarity with fusion is a plus but not required.

Key responsibilities

• Build, maintain and run pipelines that turn experimental fusion data into high-quality, consistent, documented and versioned datasets.
• Develop ML-assisted labelling and event detection (e.g. disruptions, MHD activity, diagnostic faults) using weak supervision, active learning and anomaly detection.
• Own data quality and provisioning: provenance, calibration metadata and automated validation, and manage access for the rest of the team.
• Work with physicists validating datasets and labels, and with computational scientists, control engineers and customers to integrate their data.

Required profile

• Hands-on experience building data pipelines for large scientific or industrial time-series data formats (e.g. HDF5, Parquet, Zarr).
• Experience applying ML techniques to time series in Python (PyTorch or JAX): sequence models, anomaly and event detection, semi- or weakly-supervised labelling.
• Solid software engineering habits: testing, reproducibility and data versioning.
• Evidence of independent work at a high level, such as a PhD thesis, published code used by others, or an industry project you owned; able to design data architectures independently from day 1.
• An advantage: human-in-the-loop labelling tools, pipeline orchestration (e.g. Dagster, Airflow, Prefect) and data versioning (e.g. DVC, lakeFS), and exposure to fusion data systems (MDSplus, IMAS).

Required skills

• SQL and noSQL databases, with familiarity with time-series databases such as TimescaleDB and InfluxDB.
• Python with PyTorch or JAX.

What we offer

• A key role at a fast-growing startup, shaping technical solutions and products from the ground up.
• Working alongside leading plasma and control experts on real-time plasma monitoring and control.
• Competitive salary and equity matched to seniority, plus benefits including immigration assistance.