Project Overview
This PhD project will develop a secure distributed machine learning platform to support adaptive autonomous systems in mission-critical and high-risk environments. This project requires innovative methods that allow fleets of heterogeneous agents (for example drones, vehicles, or sensors) to learn from each other, update their models, and coordinate decisions in real time while operating over unreliable networks and under potential threats. The work sits at the intersection of machine learning, distributed systems, and cybersecurity, and will be validated using realistic simulations and cloud/edge testbeds.
Aim
The PhD project aims to develop a scalable, secure distributed machine learning and AI orchestration framework that enables autonomous systems (such as drone swarms) to adapt their models and decision-making in real time under resource, connectivity, and adversarial constraints.
Objectives
- Design distributed learning methods that enable heterogeneous autonomous systems (for example drones, vehicles, and sensor nodes) to collaboratively train and update models under strict compute, energy, and bandwidth constraints.
- Develop an AI orchestration layer that can allocate tasks, models, and data across many agents, supporting real-time coordination even when connectivity is intermittent or degraded.
- Build and validate on realistic testbeds, using high-fidelity simulations and cloud/edge infrastructure to evaluate scalability, accuracy, latency, and resilience in maritime and critical-infrastructure style scenarios.
Significance
The project will provide a principled way to deploy trustworthy AI in mission-critical, high-risk environments where traditional centralised training and control are not viable. By improving robustness, adaptability, and coordination of autonomous systems, the outcomes have potential impact on areas such as disaster response, environmental monitoring, and protection of critical infrastructure.
Ideal Candidate
The Ideal candidate will have
- Computer Science, Software Engineering, Data Science, Electrical/Computer Engineering or a related discipline.
- Experience or strong interest in machine learning / deep learning (PyTorch or TensorFlow) and Python.
- Experience or strong interest in distributed systems, cloud computing (AWS/Azure), containerisation (Docker, Kubernetes) or cybersecurity
Scholarship
This ASCA grant scholarship provides a living stipend of $38,440 p.a. pro rata indexed, based on full-time studies, for up to a maximum of 3.5 years. This is only available to Domestic applicants.
Enquires
If this project interests you, contact A/Prof Sajib Mistry via the Expression of Interest.
For enquires please contact A/Prof Sajib Mistry via sajib.mistry@curtin.edu.au
Applications close 31 March 2026