Research Scientist

Research and evaluate machine-learning methods for satellite and aerial imagery, then help turn promising approaches into capabilities that can be used in production. The research spans geospatial statistics and Earth observation tasks such as image segmentation, object detection, and super-resolution, with scope to experiment and share findings.

• Design and conduct machine-learning experiments on satellite and aerial imagery.
• Build reproducible model training and evaluation pipelines.
• Assess results rigorously, documenting limitations as well as successful outcomes.
• Transfer promising prototypes to engineering for production development.
• Communicate findings through client materials, technical articles, or research publications where appropriate.

• A PhD or equivalent research record in machine learning, remote sensing, or geospatial science.
• Strong Python and PyTorch skills.
• Sound statistical knowledge and experimental design skills.
• Ability to write clearly about scientific methods and results.

• Experience with Earth observation foundation models, such as Clay, SAM, or Prithvi.
• Peer-reviewed publications.
• Experience training models on cloud or high-performance computing infrastructure.
• Contributions to open datasets or open-source tools.

• Contract role, available part-time or full-time.
• Canberra or remote.
• Room to experiment, publish, and contribute work in the open.