Flying the Plane While We’re Building It: AeRO’s Community of Research Infrastructure Experts

TL; DR

• Generative AI is seen as both exciting and concerning: it delivers powerful new capabilities and insights but is also a “black box,” so validation, transparency, and scientific rigour are critical.
• There’s particular excitement about AI and high-performance computing enabling new possibilities in social sciences, not just traditional STEM.
• Overall, the community is “building the plane while flying it,” rapidly experimenting with AI and data tools while continually stress-testing and improving them through scientific scrutiny.
  

At this year’s eResearch Australasia Conference, hosted by Australasian eResearch Organisations (AeRO), more than 500 delegates have come together to talk about one thing: how technology is transforming the way research is done. The conference is the premiere event for people working in eResearch and digital research infrastructure across Australia and New Zealand, from universities and national facilities to industry partners and vendors.

Macquarie Cloud Services’ own Naran McClung sat down with Luc Betbeder- Matibet, from the University of New South Wales and President of AeRO, who is also chairing this year’s conference. Their conversation captured exactly why this community matters, and why AI, data and cloud are reshaping what’s possible in science.

Luc describes the AeRO conference as the gathering place for “Australia’s research infrastructure experts”, the people who run the national and university systems for compute and data, connect instruments and edge devices, and help researchers make sense of it all with advanced tools, including AI.

At the conference, you’ll find:

• Technical teams behind national supercomputing and data centres
• University research IT and eResearch groups
• Vendors, sponsors and exhibitors providing cloud, storage, networking and software
• National collaborative research infrastructure organisations who run shared platforms and services

In Luc’s words, “it’s all the technical people that look after all the kit”, the invisible scaffolding that lets researchers push the boundaries of knowledge.

The conversation quickly turned to generative AI, and how it’s reshaping research. For the AeRO community, generative AI is both a powerful enabler, delivering new insights and capabilities that accelerate discovery, and a worrying black box, because scientists want transparency at every step of the research process.

That’s why this community is putting so much effort into back-tracing and validating AI outputs, testing and benchmarking tools against established methods, and understanding where AI is trustworthy, and where it’s not.

One of the most exciting shifts Luc highlights is happening in the social sciences. Traditionally, high-performance computing (HPC) has been associated with fields like physics, climate modelling, or genomics. But as datasets in areas like economics, education, policy and social research explode in size, social scientists are increasingly turning to the same powerful infrastructure, and, now, to AI-enhanced tools, to do their day-to-day work.

If there’s one image that sums up the AeRO community right now, it’s Luc’s metaphor of “flying the plane while you’re building it.” GenAI, new data architectures, cloud-native workflows and instrument-to-cloud pipelines are evolving in real time. Best practices are being written as they go. Yet that’s always been the nature of science and infrastructure: they build new systems; other scientists test and stress-test them, and the plane gets stronger, safer and more capable while it’s in the air.

For Macquarie Cloud Services, these conversations go to the heart of our role in the research ecosystem. Our work with universities, government and research institutions is about providing secure, high-performance cloud and data platforms, supporting hybrid and multi-cloud architectures that integrate with national and institutional systems, and helping research infrastructure teams scale reliably while keeping governance, security and cost in balance.

Talking with leaders like Luc at AeRO is how we ensure the platforms we build line up with the real-world needs of those “flying the plane” every day.

Full Interview Transcript 

Okay. I’m Naran McClung. I work for Macquarie Cloud Services. Now, we are here to talk about Caudit Cloud, but first and foremost I’m here with,  

I’m Luce. I’m from the University of New South Wales. That’s my day job, but I’m here as, president of AeRO and as the chair of our eResearch conference where there are 550 people just like me, attending conference this week. 

And what are they hoping to get out of it? I mean, they were sharing ideas. I thought the keynote was fascinating. I love seeing the stuff about Mars and the Rover, et cetera. Obviously, that was a massive data project as well. What are they all hoping to get out of this? Obviously, everybody’s gonna be individual. What do you think?  

So, this is Australia’s, research infrastructure experts. So, we are the people that look after our national and university systems that have compute, and data, and that take feeds from instruments, from edge systems, and are now working with all of our AI tools. To make sense of the new techniques that we need to do scientific discovery. So it’s all the technical people that look after all the kit. So we’ve got a combination of our vendor friends, our sponsors and our exhibitors. We’ve got all the national collaborative research infrastructure, friends and family, like our national supercomputing centres and data centres. That’s the type of community here.  

I love it. Now, tell me if I’m right here or tell me if I’m wrong, but I think generative AI has been hugely beneficial to this space. The stories that we hear, and we heard this from the keynote as well, the technology’s really working well here. I know that MIT published a paper, you probably saw it not that long ago, that talked about the use cases where AI can really lend itself. 

I feel like eResearch one of those.  

Definitely AI, and gen AI especially is one of those black boxes that we’re, both excited by and a little bit worried by. As science nerds, we want to have as few black boxes in the scientific discovery process as possible, and yet these gen AI tools keep giving us either new insights or new capabilities to do our work. 

And so back tracing, testing, validating, understanding those tools is really important to our community.  

Got it.  

We are so excited to be, using the data tools now, and especially in the social sciences. Our social scientists are finally getting a high performance computing tool to use for their day-to-day work. 

Amazing, fabulous stuff.  

And it’s all learning happening in real time as well. I love that there’s been enough promise shown already to encourage, do you to want to take it further and develop those best practices. It’s like, you know, you’re flying the plane and you’re building it at the same time, right? 

Absolutely. And but science has always been like that. Science is always other scientists throwing rocks at this flying plane picture that you created for us and we’re trying to make a better, stronger plane every single day as we are flying it. Yeah, absolutely. Right.  

I love it. Brilliant stuff. 

Well look, really looking forward to having more conversations here. Luc, thank you for coming on.  

Thanks Naran. Thanks for it. 

 Appreciate it. Thank you.