Not every AI technique is suitable for every research question. For Dorthe van Waarden, partner at data & AI company Rewire, it is therefore not simply about keeping up with technological developments, but about understanding which technology fits which challenge. This approach is also an important principle underlying the collaboration between Rewire and exquAIro, and the continuous development of the curriculum.
Within Rewire, Van Waarden is responsible for several teams working on AI solutions and building AI capabilities for organisations and sectors across a broad range of fields. “My role at Rewire is no longer about programming myself, but I remain closely involved in the conceptual development and structuring of AI challenges,” she explains. She is also jointly responsible for Rewire’s professional education academy, GAIn. This combination of content, technology and education closely connects to her involvement with exquAIro.
Rewire and the UMCG have been collaborating on research at the intersection of AI and biomedicine since 2018. What started with exploring how data and AI could contribute to areas including asthma research developed into a long-term collaboration and ultimately the creation of exquAIro. Today, all Dutch university medical centres are involved in the programme.
Using data to address fundamental challenges
For Van Waarden, it is precisely the combination of commercial AI expertise and medical and scientific knowledge that makes the collaboration so valuable. “I really like seeing how data can make a difference in fundamental research. We’re not asking: how can we make more profit? We’re really asking how we can solve fundamental health challenges by using technology.”
The collaboration also benefits Rewire beyond its contribution to research. The company learns from the complex challenges encountered in the medical field. “The data environments are much more complex than what we normally encounter with clients. The first time we received DNA data for a research project, we were dealing with a completely different type of data. That creates interesting challenges from a technical perspective and also drives innovation in our own techniques.”
A curriculum that keeps evolving
Van Waarden sees that same dynamic reflected in the exquAIro curriculum. Rewire brings many years of experience in professional education to the programme. Some fundamentals remain unchanged: being able to break down a research question, understand data and communicate effectively. The technical skills, however, continue to evolve.
“In the past, there was much more focus on big data. Those hard skills have really changed. That is why you have to continuously refine the curriculum based on new developments.”
That does not mean every new AI hype automatically finds its way into the bootcamp. Especially in biomedical research, it is important to consider which technology is actually appropriate for a particular application. “Innovations are valuable, but that does not mean that what already exists has become less valuable,” says Van Waarden. In biomedical research, she argues, it remains important to assess which technology best fits a specific application. The capabilities of deep learning models, for example, remain highly relevant, while generative AI is not automatically the most suitable solution for every challenge.
One development that is explicitly on the agenda is agentic AI. The topic will be introduced in the upcoming bootcamps and is expected to evolve further. This is not only about the technical capabilities of AI agents, but also about the knowledge they need in order to function effectively.
“Agents can work well if they have access to the right knowledge. That means knowledge within organisations is becoming increasingly important. It is not only structured data that needs to be made accessible, but unstructured knowledge as well.”
This also puts data security even more firmly on the agenda. Particularly in healthcare, where sensitive personal data is involved, AI systems need to operate within clear boundaries. “Risks that already exist can become larger when you use AI at scale.”
Developing together, each from their own strengths
The exquAIro curriculum is therefore not developed by a single organisation or individual. An editorial board oversees its quality, while different partners contribute their own expertise. According to Van Waarden, that is one of the programme’s strengths.
“We complement each other well. UMCG experts provide training based on their medical expertise, while Rewire mainly contributes its AI and data expertise. Together, we determine which topics are relevant and who is best positioned to take responsibility for them.”
This approach to collaboration leads to improvements in every bootcamp, according to Van Waarden: sometimes small refinements to existing training, sometimes entirely new courses.
From training to real-world impact
For Van Waarden, the next step is therefore not only further growth of exquAIro, but also a stronger connection to practical application. During the bootcamp, participants work on their own research proposals for an exquAIro Pilot Grant. The winner can subsequently receive additional support in the form of a science trainee from Rewire to help take the proposal further.
“I like that because it directly connects the training to practice. It is not just about writing a proposal; we actually help someone move forward.”
Her ambition for the coming years is clear: greater reach, both across institutions and among different types of professionals and levels of seniority. And above all, a network in which participants and organisations learn from one another. “If exquAIro can create a movement in which the whole is greater than the sum of its parts, and people genuinely inspire one another, we can take research to the next level.”