PhD Position: Sustainability of Scalable AI

This PhD project focuses on developing methods to assess the sustainability of nearly 100 nationwide AI systems within the National Lab on Education and AI (NOLAI). You will create methods to predict energy consumption, create energy labels for algorithm scalability, and guide implementers in choosing more efficient algorithms. Ready to make AI more sustainable? Apply now!

The goal of your PhD project is to develop methods to steer developments of large AI systems in such a way that they are environmentally sustainable. To this end, different designs of AI systems should be assessed during the design phase. Data flow diagrams are already used in NOLAI, and capture all the processing, storing and transmission of data: elements in the environmental impact of IT systems. You can extend these data flow diagrams with the expected environmental impact so different design variants can be considered by the team working on these systems.

You will measure and fill the unknowns uncovered in such a data flow diagram. The scalability of the core algorithms of a new nationwide AI system can be predicted using generated data sets of different sizes and measuring the environmental impact. This impact can be measured and calculated using our Software Energy Lab, which has multiple test machines with GPUs and, in the future, AI accelerators.

Development teams currently lack guidance on how to create sustainable systems. You will develop a method to help development teams choose the right algorithm and right hardware. This can be done by measuring (during the above-mentioned experiments) how the algorithms are constrained. They can be constrained by either compute power or memory bandwidth. This information can be used to calculate the theoretical maximum energy efficiency of an algorithm that is run on an architecture/accelerator. To make testing multiple architectures easier, you will leverage our existing approach to generate code from a single source code for multiple architectures and accelerators, called SaC.

You will create a way to disseminate the scalability and environmental impact within NOLAI. To this end, sustainable scalability labels of core algorithms should be made available within NOLAI. These labels can be used to design new systems. What should be included in these labels is part of this PhD research project. You will collaborate with scientific staff and PhD candidates from different disciplines and with a small team of software developers working on AI prototypes and infrastructure.

There is no teaching load in this position.

Would you like to learn more about what it’s like to pursue a PhD at Radboud University? Visit the page about working as a PhD candidate.

Requirements:

• You hold a Master's degree in Computing Science or Information Science (or a closely related field) and have knowledge of and an interest in security and privacy.
• You are eager to work in a multidisciplinary setting like NOLAI, together with AI experts, social scientists, schools and commercial partners.
• You are interested in research in an applied setting, making sure the contributions are practical and sustainable.
• You have a good command of English and good writing skills.

Salary Benefits:

• We will give you a temporary employment contract (1.0 FTE) of 1.5 years, after which your performance will be evaluated. If the evaluation is positive, your contract will be extended by 2.5 years (4-year contract).
• You will receive a starting salary of €2,901 gross per month based on a 38-hour working week, which will increase to €3,707 in the fourth year (salary scale P).
• You will receive an 8% holiday allowance and an 8,3% end-of-year bonus.
• You will receive extra days off. With full-time employment, you can choose between 30 or 41 days of annual leave instead of the statutory 20.

Work and science require good employment practices. This is reflected in Radboud University's primary and secondary employment conditions. You can make arrangements for the best possible work-life balance with flexible working hours, various leave arrangements and working from home. You are also able to compose part of your employment conditions yourself, for example, exchange income for extra leave days and receive a reimbursement for your sports subscription. And of course, we offer a good pension plan. You are given plenty of room and responsibility to develop your talents and realise your ambitions. Therefore, we provide various training and development schemes.

38 hours per week

Houtlaan 4