Role Overview
Aeolus is building a next-generation data assimilation (DA) framework in Julia that interfaces with our open-source Earth system modeling ecosystem, NumericalEarth. We are looking for a Research Scientist who can work within the Data Assimilation team to operationalize that framework.
Their primary responsibility will be to set up robust workflows to generate skillful, calibrated forecasts by assimilating multispectral observations from both public and proprietary sources. The Research Scientist will also work closely with the Modeling and Software teams to build cases that exercise new DA features and adversarially test the forecasting system.
While the initial focus is convective-scale atmospheric DA, our goal is to eventually enable fast and accurate coupled data assimilation across all components of the Earth system.
What you'll do
- Operationalize our DA framework — transform research prototypes into reliable, cycling workflows by incorporating bias correction for observations, implementing hybrid ensemble-variational assimilation, and performing forecast verification and calibration.
- Scale existing cases with new datasets, such as multispectral satellite radiances, GNSS-RO, microwave sounders, bespoke retrievals, and new algorithms.
- Develop new workflows as model components come online, with a particular focus on challenging convective-scale phenomena, including thunderstorms, tornadoes, and tropical cyclones.
- Integrate ML-based DA algorithms and help define the interfaces and benchmarks that allow us to evaluate them against established baselines.
- Build demonstration and stress-test cases that highlight new DA capabilities while exposing gaps in modeling and observation quality-control infrastructure.
Must-have qualifications
- Ph.D. in meteorology, atmospheric physics, applied mathematics, or a relevant engineering discipline.
- Expertise in operational DA and/or a strong track record of setting up quasi-realistic DA cases with WRFDA, JEDI, or a similar framework.
- Experience bridging remote-sensing observations and models, from observation operators to quality control.
- Working knowledge of DA algorithms and their practical failure modes.
- Fluency in Fortran and Python.
Nice-to-haves
- Experience with the Julia programming language.
- Experience with coupled DA, or a clear interest in moving toward it.
- Familiarity with differentiable/adjoint-based methods and ML-based DA.
- Enthusiasm for working with LLMs and transforming legacy scientific approaches with agentic AI. We expect this to change how DA systems are built, and we want someone excited to be at that frontier.
How we work
We are a small team that iterates quickly, with a fast feedback loop between scientists and engineers. The environment is fast-paced, with substantial room for personal and technical growth as the forecasting system's capabilities—and your scope—expand.
Intangibles
We are looking for someone who has demonstrated that they can take ownership of a project end-to-end and is excited about building the next generation of data assimilation systems for forecasting and control.
Who we are
Aeolus builds hardware and software to reduce weather-related risk. We're focused on mitigating both the financial and physical-world impacts of extreme weather. Our team, which includes veterans of NOAA, NASA, and NCAR, is developing first-in-class weather models and hurricane mitigation technologies in-house.
Compensation
We offer a competitive, experience-based salary and comprehensive healthcare benefits.