Energy-Efficient Decentralised Training Frameworks for Large-Scale AI Models on Geo-Distributed Infrastructure

PhD Scholarship - Energy-Effective Decentralised Training Frameworks for Large-Scale AI Models on Geo-Distributed Infrastructure

Job No.:696617

Location: Clayton campus

Employment Type: Full-time

Duration: The scholarship may be held for up to 3.5 years (fulltime) for Research Doctorate (PhD) studies

Remuneration: The successful applicant will receive

- A Research Living Allowance, at current value of $37,145 AUD per annum for PhD (2026 rate with annual indexation)
- Faculty of Information Technology Tuition Fee Scholarship (for international students only)
- Top-up scholarship of $10,000 per annum
- FIT Candidature Funding of $4,000 for the duration of the candidature
- Up to $1,265 from Monash Graduate Research Office as a one-off travel grant
- Top-up government scholarship $7,135 per annum
- Travel support of up to $2,000 per annum for first author publications to top-tier venues, provided by Pluralis

The Opportunity

This is an outstanding opportunity for a highly motivated PhD candidate interested in energy-efficient decentralized training, distributed systems, and large-scale foundation models. The successful candidate will be supervised by Dr Mohammad Goudarzi at Monash University, with co-supervision and support from leading academic and industry experts.

The candidate will join the Faculty of Information Technology at Monash University and work closely with Pluralis Research, the industry partner on this project. The project focuses on developing energy-efficient decentralized orchestration mechanisms and algorithms for training large-scale foundation models across geo-distributed infrastructure.

This research addresses a critical challenge in modern AI: how to train increasingly large models in a more scalable, accessible, and sustainable way. As AI systems continue to grow, centralized training infrastructure creates significant barriers related to cost, energy consumption, infrastructure access, and environmental impact. This project will explore new decentralized training frameworks that can better utilise distributed computing resources while reducing energy overheads and supporting more sustainable AI infrastructure.

The successful candidate will have the opportunity to work on real-world industry problems, access advanced computing infrastructure, collaborate with researchers and engineers at Pluralis Research, and contribute to high-impact research outputs, open-source tools, and potential commercial translation pathways.



The project is especially suited to candidates with strong interests in distributed systems, systems for AI/ML, machine learning systems, decentralised training, large-scale AI model training, resource orchestration, cloud/edge computing, and sustainable AI.

Through this National Industry PhD project, the candidate will receive rigorous academic training, meaningful industry experience, and professional development support. They will be embedded with Pluralis Research for part of their candidature, gaining direct exposure to industry-scale decentralised AI training platforms and practical deployment challenges. Upon completion, the candidate will be well positioned for a leading career in academia, industry research, or advanced AI infrastructure development.

To be considered for this opportunity you should fulfil the eligibility requirements listed below. The academic qualification requirements for this PhD is:

- A bachelor’s degree of at least four years in a relevant discipline, which includes a research thesis or project, with a minimum overall average grade of an honours degree equivalent to the First Class Honours; or
- A master's degree in a relevant discipline which includes a research thesis or project equivalent to at least 25 percent of one year of full-time study, with a minimum overall average grade of honours equivalent to the First Class Honours; or
- A qualification, or combination of qualifications and relevant professional experience, deemed equivalent by the GRC (or delegate).

For this particular position, applicants must also have an undergraduate or postgraduate qualification in computer science, information technology, machine learning, artificial intelligence, software engineering, or a closely related discipline. Ideally, applicants will have training and research experience in one or more of the following areas: distributed systems, systems for AI/ML, machine learning systems, decentralised training, large-scale AI model training, resource orchestration, cloud/edge computing, high-performance computing, or energy-efficient computing.

Monash University strongly advocates diversity, equality, fairness and openness. We fully support the gender equity principles of the Athena SWAN Charter.

The Project





We invite applications from outstanding PhD candidates with an undergraduate or postgraduate qualification in computer science, information technology, artificial intelligence, machine learning, software engineering, computer/electrical engineering, or a closely related discipline.

This PhD project is part of a NATIONAL INDUSTRY PHD project at Monash University, in collaboration with industry partner Pluralis Research. The project aims to develop energy-productive decentralized orchestration mechanisms and algorithms for training large-scale foundation models across geo-distributed infrastructure.

Large-scale AI training is increasingly limited by high energy consumption, infrastructure cost, communication overhead, and reliance on centralised datacentres. This project will investigate how decentralized and geo-distributed computing resources can be orchestrated more efficiently to support scalable, sustainable, and accessible training of large AI models.

The project will focus on one or more of the following research objectives:

- Develop new decentralised orchestration mechanisms for large-scale AI model training across heterogeneous and geo-distributed infrastructure;
- Design energy-aware scheduling, resource allocation, and workload placement algorithms for distributed AI training;
- Improve the communication efficiency, scalability, and reliability of decentralised training frameworks;
- Evaluate decentralised AI training systems using real-world workloads, GPU infrastructure, and industry-relevant deployment scenarios; and
- Generate open-source frameworks, algorithms, benchmarks, and research outputs that support sustainable and scalable AI infrastructure.

The project is expected to generate new knowledge, tools, and practical frameworks for reducing the energy footprint of large-scale AI training while improving the accessibility and efficiency of AI infrastructure. Expected outcomes include novel decentralised training algorithms, energy-productive orchestration techniques, open-source software artefacts, high-quality research publications, and potential translation pathways into industry systems.

Applications Close: Sunday 30 August 2026, 11:55pm AEST

We will begin the interview process as soon as suitable applications are received, so applicants are encouraged to apply early. We will not wait until the closing date to start shortlisting and interviews.

Supporting a diverse workforce

Monash University recognises that its Australian campuses are located on the unceded lands of the people of the Kulin nations, and pays its respects to their elders, past and present.

#J-18808-Ljbffr

📌 Energy-Efficient Decentralised Training Frameworks for Large-Scale AI Models on Geo-Distributed Infrastructure (Moe)
🏢 Monash University
📍 Moe