By Juan Marcelo Parra Ullauri, PhD, BT Group
That, to me, is the essence of the JOINER platform, what it enables, and what I’m trying to achieve at BT with my research into AI and Networks.
Where much of the innovation around AI in telecoms has been around how AI can be applied to the network, this work revolves around a different question:
How can we make the network a platform for AI?
Being part of the JOINER ecosystem, working with academic and industry partners through the platform, helps us explore this question in a realistic network environment.
What is JOINER?
JOINER is an international experimentation platform for telecoms innovation. In short: an intelligent network that connects over 15 of the world’s top university research labs in a real-world test environment.
Going back to my opening sentence, for us, it’s the perfect platform to quickly test,and learn, at scale, allowing us to move through the prototyping stage much faster so we can prove something works in conditions that closely resemble real networks.
That means we can significantly reduce the time it takes to get stuff out of the lab and into the world as a service, both for our customers and internally within BT.
Why JOINER?
The JOINER platform revolves around three core pillars, all of which support the work we’re doing at BT in different ways:
- Experimentation
- Collaboration
- Innovation
Let’s start with the first point.
Experimentation
For me, working in applied research, experimentation is key. It’s important to be able to showcase to the industry that new innovations can work in a real environment.
LLMs are everywhere, from chatbots to agents. These applications require a flexible infrastructure that adapts with them. Our experiment focuses on dynamic resource allocation and service activation for distributed LLM inference.
JOINER gives us a safe and controlled environment in which to do that. Ideas have the space to grow and evolve, and we can test them at a scale we could never achieve in a single lab. That scale lets us validate concepts faster and move forward with much greater confidence.
We’re not trying to reach a finished product at this stage. What we can achieve is a strong proof‑of‑concept that runs across a national‑scale environment, which is a big stepping stone for us.
Collaboration
Collaboration is one of JOINER’s biggest strengths. The platform brings together the scale and practical focus of industry with the technical depth and creativity of leading academic institutions, including those participating in the Federated Telecom Hubs.
For BT, this has been essential. Through JOINER’s research ecosystem, we gained visibility into who else was exploring similar areas, which led us to the Smart Internet Lab at the University of Bristol. They are now part of the project, and we meet with them weekly to prepare the demo.
As the experiment matured, interest grew across BT and the wider industry, helping us bring NVIDIA into the collaboration, adding deep experience in AI infrastructure to the mix. Being able to show that we were moving beyond lab-based testing and into a real network environment made a big difference in forming these collaborations.
Innovation
Innovation relies on being able to test ideas quickly and understand whether they hold up in real conditions. JOINER gives us the space to do exactly that. We can spark an idea, try it out, adapt it, and see how it behaves in a realistic network environment.
This kind of rapid validation is essential for increasing the TRL of our proof‑of‑concepts. It helps us distinguish between ideas that simply sound good and ideas that actually work when deployed at scale. And when something doesn’t work, we can pivot fast, adjust the approach and test again without losing momentum.
All of this means we can innovate faster and with more confidence. It benefits our research, our customers and the wider organisations involved. And more broadly, it helps the UK strengthen its position in the global telecoms landscape.
Looking ahead
From the experiment side, our focus now is to validate this first set of ideas on scaling LLM inference in a realistic telecom environment. After that, we want to explore new research areas, including how the network itself might contribute to making AI more efficient, sovereign and easier to operate. JOINER will be central to this work, connecting research labs across the UK, giving access to emerging technologies and acting as a catalyst for collaboration.
But there are so many more use cases to explore. And many more potential partners we could collaborate with through the platform. So, I’ll end by saying this:
If you have an idea that could work in telecoms, and you need a safe environment in which to test and develop that idea in partnership with some of the greatest minds on the planet, come talk to JOINER ecosystem.
