Csiro Postdoctoral Fellowship In Imaging Black Holes

CSIRO acknowledges the Traditional Owners of the land, sea and waters, of the area that we live and work on across Australia. We acknowledge their continuing connection to their culture and pay our respects to their Elders past and present. View our vision towards reconciliation

• Child Safety

CSIRO is committed to the safety and wellbeing of all children and young people involved in our activities and programs. View our Child Safe Policy.

• The Opportunity
• Do you have a PhD in Computer Science, Astrophysics / Physics, Mathematics, Electrical Engineering, Software Engineering or a similar discipline?
• Does the challenge of making images of black holes and peering close in on them to understand foundational physics thrill you?
• Does providing novel solutions to image classification and reconstruction problems using Machine Learning / Deep Learning sound exciting to you?

CSIRO Early Research Career (CERC) Postdoctoral Fellowships provide opportunities to scientists and engineers who have completed their doctorate and have less than three years of relevant postdoctoral work experience. These fellowships aim to develop the next generation of future leaders of the innovation system.

The CERC Fellow will develop new and independent image reconstruction methods based on deep learning methods and the theory of interferometry to determine the spacetime warping and morphology around the event horizons of black holes. A planet-sized network of telescope arrays using Very Long Baseline Interferometry (VLBI) is needed to image close in on the event horizons of supermassive black holes, which is extremely challenging because the telescopes must be accurately calibrated for delays and attenuations in the signal paths, and the spatial information sampled by the telescopes contains only sparse and incomplete information. The independent end-to-end image reconstruction technique addressing these challenges will be applied to real data obtained from VLBI networks, including primarily the recent and famous example of the supermassive black holes at the centres of our Milky Way and M87 galaxies by the Event Horizon Telescope (EHT).

This independent approach to interferometric imaging will provide a powerful constraint on the structure of spacetime warping near black holes. It will be instrumental in testing fundamental physics such as the theory of General Relativity and will also reveal the relativistic astrophysics at the black hole site. In parallel, this project will benefit both radio astronomy and deep learning, computational astrophysics in general, as well as interferometric imaging problems in other domains such as medical imaging and seismology.

The CERC Fellow will work in close collaboration with the principal investigator Dr. Nithyanandan Thyagarajan (Space & Astronomy), and co-principal investigators Dr. Ivy Wong (Space & Astronomy) and Dr. Foivos Diakogiannis (Data61) and have the opportunity to utilise state-of-the-art GPU infrastructure both at CSIRO and National facilities like Pawsey Supercomputing Centre's High Performance Computing facility, Setonix, for this project, giving a competitive advantage.

• Your duties will include
• Developing and implementing machine learning / deep learning solutions to interferometric imaging problems where sparsity of information and precise calibration of the data collected by the interferometric elements are the primary limitations to image reconstruction quality.
• Effectively translating machine / deep learning techniques to interferometric imaging problems in radio astronomy using CSIRO high-performance computing facilities.
• Effectively translating machine / deep learning techniques from radio interferometry to other fields that use interferometry for imaging such as medical imaging, seismology, etc.
• Publishing results and methods in relevant international journals (and conferences) and releasing solutions onto public repositories such as GitHub.
• Location
• Kensington, Western Australia
• Salary
• AU$92k - AU$101k plus up to 15.4% superannuation
• Tenure
• Specified term of 3 years
• Reference
• 93599
• To be considered you will need
• A doctorate (or will shortly satisfy the requirements of a PhD) in a relevant discipline area, such as Computer Science, Astrophysics / Physics, Mathematics, Electrical Engineering, Software Engineering or similar. Please note : To be_ _eligible for this role you must have no more than 3 years (full-time equivalent) of postdoctoral research experience_._
• Demonstrated expertise in designing and implementing ML frameworks and workflows.
• Demonstrated expertise in high-level programming language(s) such as Python or similar.
• High level written and oral communication skills in English with the ability to represent the research team effectively internally and externally, including the presentation of research outcomes at national and international conferences.