Machine Learning Engineer

What we do and why it matters

Terrafuse is building AI technology that accelerates climate physics models from hours on a supercomputer to seconds in the cloud, greatly improving our ability to protect against disastrous impacts of climate change and plan renewable energy infrastructure.

Climate change is projected to displace hundreds of millions of people and cause trillions of dollars in damage by 2050. Protecting people and infrastructure against the damage is only possible with accurate and granular risk models, because broadly protecting all areas is too expensive. For example, increased frequency of wildfires is forcing insurers out of California - however, with our product that accurately quantifies wildfire risk property by property, it becomes economical for them to continue providing coverage.

Additionally, our wind product helps optimize global wind energy production, which is growing 20% year over year as the world transitions to a clean energy economy.

Our technology combines our academic expertise in climate science and machine learning, developed from our work at Lawrence Berkeley National Lab and supported by the National Science Foundation, the U.S. Air Force, and Microsoft AI for Earth. Our customers and stakeholders include government, insurance, aviation, and clean energy industries. 

The position

At Terrafuse our vision is to build a state-of-the-art climate platform powered by AI. We are seeking someone passionate about climate and experienced with deep learning models and architectures to join our research and engineering team. This role reports to the CTO and the scope includes: 

1. Deep learning model development and implementation. Develop deep learning models for our wildfire and wind products with a focus on model implementation in production environments. 
   • Develop and implement spatiotemporal, probabilistic models for wildfire risk. 
   • Develop and implement real-time meteorological forecasts of wind and other variables; help scale and benchmark ML-based forecast models against existing HPC numerical weather prediction models. 
2. Machine learning infrastructure. Work with the CTO and Lead Software Engineer to develop a cloud-native infrastructure for ML model development and deployment as we work toward continuous integration and delivery of ML products. Help spec and implement replicable ML pipelines that span data and model training, validation, and inference. Work with the Lead Engineer to integrate the ML pipeline with the backend infrastructure to support model deployment in production environments.
3. Contribute to team building. Help recruit new technical talent as we grow.

Traits we’re looking for

• Expertise with deep learning models and architectures in a research environment, including convolutional neural networks, generative models, self-supervised learning, time-series methods and probabilistic models 
• Experience with the Python scientific computing stack (Numpy/SciPy/Xarray) and one or more deep learning frameworks such as PyTorch (which we use) or Tensorflow
• Degree (Ph.D. preferred) in physics, applied math, engineering or climate/meteorology/oceanography
• Experience with scientific computing and/or high-performance computing a plus
• Curiosity and enthusiasm for independent research; the ability to learn and implement new modeling approaches from the ML literature; and the ability to validate and iterate modeling approaches to achieve target metrics and deliver on project scope

What we offer:

• You will work on challenging problems of critical importance with a team of climate and machine learning scientists with academic research backgrounds
• You will have major impact over our technology, product and company culture 
• We offer a competitive salary and equity compensation, commensurate with the importance of the position and the experience level required

Terrafuse is committed to creating an inclusive employee experience for all. Regardless of race, gender, religion, sexual orientation, age, disability, or parental responsibilities, we firmly believe that our work is at its best when everyone feels free to be their most authentic self.