Spire Global is seeking an applied mathematician scientist that will contribute to the Company’s effort of developing its statistical models’ portfolio and advanced Data Assimilation (DA) methods supporting our satellite missions. This is an exciting opportunity for motivated scientists to make a difference through state-of-the-art predictive modeling, statistical models, machine learning and data assimilation techniques using the available surface observations and large amount of Radio Occultation (GNSS-RO), Reflectometry (GNSS-R) and other satellite data.
The successful candidate will join the Spire Office in Luxembourg and will report to the Statistics Team under the Global Validation Model branch at Spire Office in Boulder, Colorado, USA. The selected candidate will have opportunities to work with top-level scientists at Spire as well as scientists from around the world on issues that matter.
The successful candidate will work with Spire DA and modeling team members to implement and evaluate statistical inference methods, statistical and machine learning models, and DA methods and make use of large datasets of surface and satellite observations.
Responsibilities will include the following tasks:
Propose and implement innovative predictive modeling and Bayesian inference approaches to increase the forecast skills of weather fields distributions.
Propose and implement hybrid techniques combining data assimilation and machine learning strategies to improve the impact of large sets of GNSS-RO, GNSS-R and other satellite data to the weather forecast skills.
Propose and implement sensitivity analysis methods to quantify the impact of GNSS-RO data into the forecasts.
Propose and implement smooth techniques for non-differential physical processes included in the numerical weather prediction models or forward observation operators to allow the use of adjoint and automatic differentiation techniques inside the predictive models.
Working with meteorological data sets from various sources.
Working with the software engineering team to define most effective software solutions.
Presenting research findings at scientific conferences or workshops.
Applicants must have either a MS or PhD degree in Applied Mathematics, Computer Science, Meteorology, or Atmospheric Science or equivalent working experience in advanced statistical inference methods, machine learning algorithms and/or DA techniques.
Working experience with combining multiple models and ensembles for predictive modeling and uncertainty quantification (Mixture models for density estimations, Bayesian Modeling Averaging, etc.).
Working experience with numerical optimization and automatic differentiation techniques.
Working experience with statistical inference methods for uncertainty quantification such as Bayesian inference and Adaptive Metropolis algorithms.
Working experience with machine learning techniques including linear and non-linear regression techniques, classification algorithms and model assessment and selection.
Working experience with cutting-edge DA methods, such as 3D-Var, 4D-Var, ensemble Kalman filter and hybrid ensemble-variational methods.
Working experience with satellite data and DA systems involving complex NWP models, such as WRF, GFS and ECMWF forecast models.
Working knowledge of Fortran-90, Python, Matlab, Linux scripting and code management practices.
Demonstration of enthusiasm and ability to work in a development team that never stops improving predicting modeling techniques, statistical and DA methods and their outcomes.