👋🏻
I translate scientific breakthroughs into high-scale industrial systems. Across 7 years as an expert in AI and Cybersecurity, I have focused on discovering novel math, funding multi-million dollar R&D programs, building algorithmic engines for 50M+ users, shipping production software for the Australian Defence Force, and governing global infrastructure through WIPO and the QUAD summit.
Beyond the University of Melbourne, I am the Chief Technology Officer of MetaGen and the owner of Prismore AI. I also serve as a founding board member of Music Technology Australia.
Contact me here to discuss technical leadership, algorithmic revamps at scale, or AI governance.
About Me
A technical leader and AI architect specializing in high-scale algorithmic revamps and mathematically robust system design. I bridge the gap between theoretical breakthroughs in Certified Robustness and Dynamical Systems and the practical realities of production deployment.
- LocationMelbourne, Australia
- Age36
- NationalityAustralian
- InterestsCycling, Running, Weightlifting, Formula 1, Weird Fiction
- StudyMonash University
- EmploymentUniversity of Melbourne
Education
Doctor of Philosophy (Science)
Developed fast, scalable numerical techniques for nonlinear systems. Expert in Certified Robustness, Game Theory, and Dynamical Systems.
B. Aerospace Engineering & B. Science
Optimized complex structural equations using Geometric Programming and Genetic Algorithms, achieving 50% mass reduction for aerospace applications.
Work
Owner
Leading an early-stage startup providing white-label AI safety platforms for healthcare and high-security industries, preventing unauthorized confidential data leakage.
Chief Technology Officer
Directing the technical roadmap and orchestrating cross-functional delivery with international contractors to build metadata management systems for the global music industry.
Founding Board Member
Representing technical and research interests on the board of Australia's peak industry body for music technology; driving ethical technology standards and advocating for artist-centric technical infrastructure.
Senior Research Fellow
Directed $3.5M in R&D funds across multi-year programs. Produced 40+ peer-reviewed papers (NeurIPS, ICML, ICLR) and 30+ technical reports. Transformed research into production software for the Australian Defence Force specializing in signals intelligence and AI cybersecurity. Currently managing technical delivery for 17 researchers.
Consultant Mathematician & Data Scientist
Led core product revamp for 50M+ user platform; redesigned bid-ranking algorithms for performance and transparency. Trace millions in lost revenue for ASX-listed clients via quantitative analysis.
Principal – Quantitative Trading & Risk Management
Developed proprietary mathematical models for high-stakes variable price markets with an annual turnover of $2 million, focusing on active capital allocation and risk management.
Media Training
Specialized training for translating complex AI and Cybersecurity topics for TV, Radio, and Press.
Skills
Projects
Music Technology Australia
50M+ User Bid-Ranking Revamp
Mission-Critical ADF Software
Executive Training on AI
Prismore AI
Abridged Bibliography
- Cullen, Andrew Craig, Montague, Paul, Erfani, Sarah Monazam and Rubinstein, Benjamin IP. "Position: Certified Robustness Does Not (Yet) Imply Model Security". In: Forty-second International Conference on Machine Learning (Position Paper Track) (2025).[]
- Cullen, Andrew C, Rubinstein, Benjamin IP, Kandeepan, Sithamparanathan, Flower, Barry and Leong, Philip HW. "Predicting dynamic spectrum allocation: a review covering simulation, modelling, and prediction". In: Artificial Intelligence Review (2023).
- Cullen, Andrew C, Montague, Paul, Liu, Shijie, Erfani, Sarah M and Rubinstein, Benjamin IP. "Double Bubble, Toil and Trouble: Enhancing Certified Robustness through Transitivity". In: Advances in Neural Information Processing Systems (Neurips) (2022).[Selected for Oral Presentation (Top 1.7% of papers) ; arXiv]
- Cullen, Andrew C and Clarke, Simon R. "A fast, spectrally accurate homotopy based numerical method for solving nonlinear differential equations". In: Journal of Computational Physics (2019).
Recent Posts
Get in touch.
Reach out to discuss AI technical leadership, algorithmic redesigns, or R&D strategy.
- Location
- Melbourne, Australia
- Github
- andrew-cullen
- Andrew Cullen