Our client is an early-stage, venture-backed deep-tech company developing next-generation tools for subsurface characterization to accelerate clean energy deployment. Their work sits at the intersection of numerical physics, geoscience, and advanced machine learning, with a specific focus on reducing the cost and uncertainty of geothermal exploration.
Founded by experts in physics and computation, the team is intentionally small, highly technical, and academically rigorous. They value first-principles thinking, intellectual curiosity, and a deep personal commitment to climate and clean energy impact. The company has over two years of runway following a recent pre-seed raise and is preparing for its next funding round.
As a Machine Learning Research Scientist, you will help build research-grade machine learning models that tightly integrate physical laws with data. You will work closely with domain experts in physics simulation and software engineering to translate geophysical insight into principled ML architectures that can be trusted in real-world energy decisions.
This is a selective, fundamentals-driven research role. Our client is not looking for a tooling-only ML profile, but for someone who thinks in mathematics and physics first.
Key Responsibilities
- Develop machine learning models grounded in mathematical and physical principles to augment numerical physics simulations
- Design and implement algorithms that explicitly incorporate differential equations and physical constraints
- Collaborate closely with physicists and engineers to translate geophysical understanding into ML architectures
- Influence the direction of core ML research within a lean, mission-driven team
- Build reproducible research workflows that feed directly into tools for clean energy deployment
- PhD or equivalent research experience in Mathematics, Physics, or a closely related quantitative field
- Strong mathematical maturity with regular use of linear algebra, differential equations, and numerical methods
- First-principles problem-solving approach rather than reliance on high-level ML abstractions
- Strong Python skills and experience writing clean, research-grade ML code
- Genuine motivation for climate, clean energy, and scientifically meaningful work
- Experience in scientific machine learning, including PINNs, operator learning, or surrogate modeling
- Background in numerical simulation or high-performance computing
- Exposure to geophysics, subsurface modeling, or energy-domain problems
- You can clearly articulate the why, how, and what of your modeling decisions, particularly where physics and ML intersect
- You produce reproducible research that improves the speed and quality of subsurface predictions
- You contribute to both foundational algorithms and practical tools used by scientists and engineers
- Video interview with the founding team
- On-site interview with the technical team over one full day
