Machine Learning Scientist

The R&D Team at Tomorrow.io is a dynamic mix of scientists and engineers. Our mission is to generate the best and most novel data and models across all times: historical, real-time, and forecast. The story just begins when the data hits our ingest and post-processing services. Every product that the user sees is the result of a pipeline of algorithms that needs to be run quickly and continuously, in an operational environment. We are the team that builds the architecture behind the data and the models, preparing the weather analyses for the Product and Engineering team to serve the masses.

We are seeking a Machine Learning Scientist with experience in AI weather prediction to help advance Tomorrow.io’s next-generation forecasting capabilities. In this role, you will combine atmospheric science expertise with state-of-the-art machine learning methods to improve forecast skill, including developing new ways to leverage observations from Tomorrow.io’s microwave sounder constellation. You will work collaboratively from proof-of-concept through deployment, with a focus on translating research advances into operational, customer-impacting products.

• Conduct research and development at the intersection of machine learning and weather prediction.

• Develop, train, evaluate, and improve AI-based weather prediction models, with a focus on measurable improvements in forecast skill and customer value.

• Develop approaches to maximize the value of observations from Tomorrow.io’s satellite constellation for weather prediction.

• Explore machine learning approaches for incorporating observations into forecast systems, including ML-based data assimilation and related methods.

• Work with large atmospheric and geophysical datasets and build reproducible, maintainable ML workflows.

• Collaborate with scientists and engineers to transition successful research from proof-of-concept into scalable, operational systems.

• Communicate results clearly through internal reviews, technical discussions, and, where appropriate, conferences and peer-reviewed publications.

• Graduate degree in atmospheric science, meteorology, computer science, or a related quantitative field.

• 2+ years of experience developing machine learning approaches for weather prediction or closely related geoscience problems. Relevant PhD research developing ML models may count toward this experience.

• Hands-on experience training, evaluating, and working with deep learning models for atmospheric science.

• Strong understanding of ML development best practices specific to atmospheric science, including experimental design, model evaluation, testing, documentation, and code review.

• Strong written and verbal communication skills and the ability to explain complex technical results to both technical and non-technical audiences. Demonstrated ability to collaborate across disciplines and deliver high-quality work.

• Experience conducting independent research, demonstrated through publications, research leadership, open-source contributions, or other technical work.

• Knowledge of or experience with data assimilation, including traditional and/or machine-learning-based approaches, is a strong plus.

• Knowledge of satellite remote sensing and experience working with satellite observations.

• Experience developing production-quality scientific or machine learning software.

• Familiarity with modern architectures used in weather and geoscience ML, including graph neural networks and transformers.

• AI-first mentality towards research and development (e.g., using AI-assisted development tools)

So, if you're looking to join a team that is not only at the forefront of