Jua develops artificial intelligence systems for weather forecasting, with a focus on applications in energy trading and renewable energy. The company's technical work centers on the Earth Physics Transformer (EPT), a model designed to learn from both historical weather data and the fundamental physics of the atmosphere. This end-to-end transformer architecture integrates observed weather data with physics-based simulations, including ERA5 reanalysis data, and is noted for its fast processing speeds.
Beyond the core model, Jua offers a Weather Intelligence Platform and associated APIs. These tools are intended to provide weather intelligence for energy markets, helping users respond more quickly to weather changes and maintain consistent forecast accuracy across different time horizons.
The company's technical domains include AI, transformer architecture, physics-based simulation, and atmospheric dynamics modeling. Its culture is oriented around the concept of building what it describes as "physical AI," a theme reflected in its roles and hiring materials.
