Machine Learning Engineer Physical Systems at Carbon Re

Join a mission-driven team applying cutting-edge AI to decarbonize the world’s heaviest-emitting industries. You will develop supervised learning models and soft sensors to optimize real-time operations in cement and steel plants. This role bridges fundamental ML research and production-grade engineering to achieve measurable, gigatonne-scale reductions in global carbon emissions.

• Build supervised learning and surrogate models for time series prediction and real-time process optimization in energy-intensive manufacturing facilities.
• Design and deploy robust MLOps pipelines using Docker and MLflow to handle noisy, large-scale industrial sensor data across multiple plants.
• Collaborate with process engineers to integrate thermodynamic constraints and conservation laws into hybrid ML models for improved reliability.

• Strong background in supervised ML for time series modeling, regression, and classification applied specifically to physical or industrial systems.
• Proven experience deploying production models with CI/CD, monitoring, and handling real-world data issues like drift and missing values.
• Expertise in the Python stack (PyTorch/TensorFlow, scikit-learn) and the ability to translate chemical or process engineering knowledge into model design.

• Tangible climate impact: Your models directly reduce millions of tonnes of CO2 emissions by optimizing high-temperature industrial processes at scale.
• Technical innovation: Work at the intersection of ML and physics, utilizing graph neural networks, state-space models, and physics-informed neural networks.
• Strong commercial traction: Deploy solutions already used by global leaders like Heidelberg Materials through a strategic partnership with industrial giant ABB.