WindLens is an early-stage AI startup incubated at EPFL, working at the intersection of machine learning, weather, and energy markets to help renewable operators and traders manage risk in a rapidly changing grid.
We are backed by the EPFL AI Center, the Swiss National Science Foundation, and Innosuisse, and work with leading European wind operators and trading desks across active pilots.
We are a small, technical team based in Lausanne. We move fast, work directly with customers on real production data, and look for people who want ownership over hard problems from day one.
The role
We are looking for a machine learning engineer/scientist to join us for an internship focused on advancing our local renewable production modelling. You will work directly with the founders on research and engineering problems that have an immediate impact on the product and on customer pilots.
Concretely, you will:
* Develop and improve machine learning models for local renewable production at the asset and portfolio level
* Explore probabilistic methods to quantify uncertainty in production estimates (ensemble approaches, quantile regression, conditional generative models, calibration techniques)
* Stay close to recent research in ML for weather and renewables, and prototype new techniques against real benchmarks
* Build and maintain a scalable codebase that handles large-volume weather, SCADA, and market data — including efficient data pipelines, reproducible training infrastructure, and model serving
* Run experiments end-to-end: from hypothesis to deployment, with rigorous evaluation against production data
What we're looking for
* Strong background in machine learning, ideally with applied experience in time series, spatiotemporal data, or probabilistic modelling
* Master's-level or PhD candidate in computer science, applied mathematics, physics, engineering, or a related field
* Comfortable working in Python (PyTorch in particular), with solid software engineering instincts — clean code, version control, reproducibility
* Experience handling large geospatial or weather datasets (xarray, NetCDF, Zarr, Parquet) is a plus
* Background or strong interest in weather forecasting and/or energy markets is highly encouraged
* Self-directed, comfortable with ambiguity, and able to drive a problem from research to production
#J-18808-Ljbffr