AI Research Scientist

Role overview Lead research and development of the machine learning model at the center of a system that learns from Earth observation and remote sensing data. You’ll work with scientific and engineering colleagues to decide what the model should learn, develop and evaluate approaches, and bring promising results into a product that helps answer questions about places and environmental change. The work spans early research experiments through large-scale training and practical integration.

Responsibilities - Assess the current model architecture and revise it when experiments and evidence support a better approach. - Work with domain experts to uncover useful patterns and semantics in large, unconventional datasets. - Define evaluation measures suited to the intended applications, rather than relying only on established benchmarks. - Test ideas with efficient small experiments, then collaborate with engineers to scale successful approaches to very large training runs. - Partner with the agent and product engineering teams on how the model is called and how its outputs can be made more useful. - Set research milestones and resource needs, identify risks, and share findings through written work and conference participation.

Requirements - A PhD in computer science, engineering or a related area, or equivalent experience. - At least two years of industry experience building and shipping AI systems. - Strong knowledge of current AI research, demonstrated through original contributions, publications or substantial projects. - Experience taking exploratory work under limited compute or data conditions into performant, large-scale training. - Ability to design models for data types beyond text.

Nice to have - Experience with image, time-series, sensor, geospatial or Earth observation data, or training models on other non-language modalities. - Background in Earth science, mathematics or physics, especially as applied to physical systems and sensor data. - Fluency in Python and PyTorch, with the ability to build efficient, reliable training and inference code.

Benefits and work setup Full-time role based in San Francisco or remote. The listing describes competitive salary, early-stage equity, and health, dental and vision coverage. The work supports applications in areas such as conservation and climate response.