Location: San Francisco
Team: Applied Science, Numerical Weather Prediction
Type: Full-time
Why this role exists
We're building a next-generation, differentiable atmospheric model for hurricane science and forecasting. One of our key goals is to build a model that's up to an order of magnitude faster than the state-of-the-art, using cutting-edge physics parameterizations and numerical methods, novel GPU-specific algorithms and performance optimizations. Our work requires expertise at the intersection of:
- atmospheric physics
- computational fluid dynamics
- differentiable programming
- data assimilation
- GPU performance engineering
A person with skills in any of these areas has the potential to contribute to this ambitious project.
What you'll do
- Develop, validate, and deploy a cutting edge coupled atmosphere-ocean model with the potential to transform weather research and forecasting
- Develop numerical methods, algorithms, core physics components, and forecasting infrastructure for two-way nesting and atmosphere-ocean coupling on single and multiple GPUs
- Advance science and engage with academics, students, and industry collaborators by contributing to the open source atmospheric modeling library Breeze.jl and the open source ocean modeling library Oceananigans.jl
- Develop differentiable workflows, data assimilation infrastructure, and ML components that build on open source Breeze capabilities for the Aeolus Forecasting System (AeFS)
- Perform benchmarking, profiling, performance optimization, and algorithm development for multi-GPU simulations, collaborating with academic and industry engineers, with the goal of developing the world’s fastest atmosphere model
Must-have qualifications
- 2+ years of experience using and developing GPU-based software for computational fluid dynamics
- Education in physics, engineering, or Earth sciences