Share this Job

Research Scientist - Efficient machine learning for time-series data

Date: 08-Sep-2022

Location: Brisbane, QLD, AU

Company: CSIRO

Acknowledgement of Country

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

The Opportunity

Our Distributed Sensing Systems (DSS) group is one of the leading large-scale sensing and analytics groups in the world, based in Brisbane, Queensland, Australia. We are part of the Cyber-Physical Research Program at CSIRO’s Data61 Business Unit. The group’s research focuses on creating integrated sensing, AI/ML, and telemetry technologies that will radically improve the cost and quality of data gathering on a large scale to enhance the understanding of our natural and built environments. 


  • An innovative and collaborative workplace with fantastic flexibility 
  • Work in a world-class research & development precinct 
  • Join CSIRO and support Australia's premier scientific organisation! 


This is a research scientist position, expected to conduct research on efficient machine learning algorithms for time-series data on edge devices and implement such algorithms on edge hardware, for projects in digital agriculture – eGrazor, digital manufacturing – FDMF/Maven, and smart building/digital twin domains. Specifically, the position will develop lightweight ML models to analyse multi-modal time-series data from multiple edge devices and sensors such as temperature/humidity, or accelerometer, to understand/classify the context in which a sensor device operates, or identify activities of the target object. 

Your duties will include:

  • Develop multi-variate/multi-modal algorithmic solutions that are suitable for distributed and in-network processing or edge computing.
  • Implement and evaluate the developed algorithms and methods efficiently using Python libraries such as scikit-learn, TensorFlow, and PyTorch.
  • Liaise with domain scientists/experts to validate algorithms and tools on the appropriate embedded systems or edge computing platforms.
  • Publish results in relevant reputable journals and conferences and prepare patent applications.
  • Recognise and exploit opportunities for innovation and the generation of new theoretical perspectives, and progress opportunities for further development or creation of new lines of research.


Location: Brisbane (Pullenvale), Queensland

Salary: AU$102,724 to $AU111,165 + up to 15.4% superannuation

Tenure: Indefinite/Specified term of x years/Casual

Reference: 88767

To be considered you will need:


  • A PhD (or an equivalent combination of qualifications and research experience) in a relevant field.
  • Solid knowledge of machine learning (preferably in statistical learning/deep learning).
  • Demonstrated experience in models simplification using techniques such as model/knowledge distillation, binarization, etc. 
  • A sound history of publication in high-rank peer reviewed journals and/or authorship of scientific papers, reports, grant applications or patents, in machine learning or systems areas.
  • The ability to work effectively as part of a multi-disciplinary, regionally dispersed research team, plus the motivation and discipline to carry out autonomous research.
  • Proficient in Python, C++ or equivalent. 
  • High level written and oral communication skills with the ability to represent the research team effectively internally and externally, including the presentation of research outcomes at national and international conferences.
  • A record of science innovation and creativity, including the ability & willingness to incorporate novel ideas and approaches into scientific investigations.



  • Experience or interest in one or more of the following: designing and implementing time-series data analysis/mining algorithms with information theory, statistics or neural networks-based models.  
  • Good experience with high-dimensional, multimodal time-series data streams. 
  • Good experience using GPU-assisted model acceleration and source code versioning systems such as Git. 
  • Experience working in object/event detection, or activity classification.


For full details about this role please view the Position Description


Applications for this position are open to Australian/New Zealand Citizens, Australian Permanent Residents or you must either hold, or be able to obtain, a valid working visa for the duration of the specified term. Appointment to this role is subject to provision of a national police check and may be subject to other security/medical/character requirements.


To enter a CSIRO site, CSIRO will require you to show proof of vaccination against COVID-19 (or show a valid medical exemption from vaccination). If you are unable to meet this requirement, then you must return a negative result on a Rapid Antigen Test (within 48 hours prior to attending site) and wear a face mask whilst on the CSIRO site. These requirements apply if you are attending a CSIRO site as part of a recruitment process.

Flexible Working Arrangements

We work flexibly at CSIRO, offering a range of options for how, when and where you work. 

Diversity and Inclusion

We are working hard to recruit people representing the diversity across our society, and ensure that all our people feel supported to do their best work and feel empowered to let their ideas flourish. 


At CSIRO Australia's national science agency, we solve the greatest challenges through innovative science and technology. We put the safety and wellbeing of our people above all else and earn trust everywhere because we only deal in facts. We collaborate widely and generously and deliver solutions with real impact. 

Join us and start creating tomorrow today!

How to Apply

Please apply on-line and provide a cover letter and CV that best demonstrate your motivation and ability to meet the requirements of this role.

Applications Close

9th October 2022, 11:00pm AEST