AI/ML Scientist, Planetary Science

About the role

This position focuses on the development and deployment of advanced machine learning systems intended to power autonomous scientific discovery for future interplanetary missions. The role owner will act as a core technical lead in creating models that can function reliably on flight hardware while operating within the harsh constraints of Mars orbit. You will serve as a primary architect for a joint initiative connecting the Relativity Interplanetary Sciences Program with the computational expertise of Polymathic AI. Much of your work will center on translating complex planetary physics into robust algorithms that run directly on spacecraft processors without human intervention. A significant portion of your contribution will involve ensuring that models remain reliable and interpretable when operating far from Earth with limited ability to debug in real time. You will be responsible for bridging the gap between high-fidelity simulation data and the noisy, limited reality encountered by sensors in deep space. Success in this role requires a mindset oriented toward both rigorous scientific inquiry and the practical engineering necessary to survive in an extraterrestrial environment.

Key facts

What you'll do

• Construct physics-informed machine learning models that simulate Martian atmospheric dynamics using terrestrial data sources to predict weather patterns for execution on spacecraft.
• Engineer a multi-modal data fusion pipeline capable of synthesizing two-dimensional imagery, three-dimensional surface topographies, geological maps, and radar depth soundings into a unified representation.
• Design and implement autonomous decision-making frameworks that allow a spacecraft to perform real-time inference and adjust its observation schedule without waiting for ground control commands.
• Lead the full lifecycle management of complex software systems, guiding a project from initial data exploration and hypothesis formation through to final deployment and performance validation.
• Partner with scientific staff at the Flatiron Institute and New York University to align computational methods with cutting-edge astrophysical research goals.
• Optimize legacy algorithms to meet the strict computational and power limitations imposed by flight hardware environments.
• Validate model outputs against physical constraints to ensure that predictions remain plausible under the extreme conditions of the Martian environment.
• Document methodologies and results in a clear and comprehensive manner to facilitate knowledge transfer and peer review within the scientific community.
• Mentor junior data scientists and engineers by providing technical guidance on best practices for machine learning deployment in aerospace contexts.
• Continuously monitor the performance of deployed systems to identify anomalies and drive iterative improvements based on real-world operational data.

Requirements

• Hold a Doctor of Philosophy degree in computer science, physics, machine learning, or a closely related technical discipline.
• Demonstrate extensive experience in applying machine learning techniques to datasets derived from physical or scientific measurement systems.
• Show advanced proficiency in transfer learning, domain adaptation, and model fine-tuning strategies specifically for scenarios involving low-data regimes and out-of-distribution conditions.
• Possess a working knowledge of multi-modal data fusion techniques and the theoretical principles underlying their integration.
• Exhibit a proven ability to own complex projects end-to-end, navigating ambiguity from initial data understanding through final production deployment.
• Maintain a strong grasp of software engineering best practices, including version control, testing, and modular code design.
• Display competence in numerical computing and scientific programming using relevant libraries and frameworks.
• Communicate effectively with both technical and non-technical stakeholders to align project objectives with mission requirements.

Skills & tools

Deep learning

Machine learning

Multi-modal data fusion

Transfer learning

Domain adaptation

Atmospheric modeling

Practical notes

This position requires approximately 10% travel to New York to work with the Polymathic AI team at the Flatiron Institute.

Compensation includes salary, equity, PTO, sick leave, parental leave, and a learning and development stipend.

Reasonable accommodations are available for applicants by contacting accommodations@relativityspace.com.