See Brightband at AMS 2025 Read more Opportunity
- Be one of the initial hires at a remote startup, started by experienced entrepreneurs, developing a transformative approach to earth system modeling.
- Build the world’s best weather forecast using a data-driven, end-to-end learned approach.
- Join a multi-disciplinary team committed to open science and sharing results with the broader weather and climate communities.
Requirements
- BS, MS, or PhD in computer science, mathematics, applied statistics, machine learning, physics, or equivalent industry experience.
- Practical experience in applying experimental ideas to real-world problems.
- Strong understanding of machine learning and statistical methods.
- Experience with Python-based ML frameworks such as PyTorch or JAX.
- Proficiency in running, tracking, and analyzing experiments, with the ability to instrument them with meaningful metrics and visualizations.
- Strong troubleshooting skills to diagnose and resolve issues in machine learning workflows.
- Ability to work independently.
- Flexibility and adaptability to work on diverse projects and pivot when necessary.
Great to Have
- Expertise in developing and optimizing data loaders for various storage solutions.
- Experience with distributed, multi-node training for machine learning models.
- Proficiency with software environment management tools such as conda or Docker.
- Familiarity with ML architectures such as Graph Neural Networks (GNNs), transformers, and diffusion models.
- Experience working with physical sensor data.
- Familiarity with the basic principles of numerical weather prediction systems.
Responsibilities
- Collaborate with the founding team to advance the state of the art in weather forecasting using a data-driven, end-to-end learned approach.
- Identify and prototype promising ML approaches from the broader research community
- Conduct experiments, analyze results, and scale up approaches that demonstrate experimental success.
- Establish best practices and workflows for distributed training to ensure efficient and effective scaling of machine learning models.
- Collaborate with our data engineering team to create efficient and maintainable data loading pipelines for a wide range of sensor data.
- Promote engineering and research best practices by conducting code reviews and ensuring high-quality code.