Internship is a win-win for Connected Data Group and University of Valencia students
Recent news.
We live data
Internship is a win-win for Connected Data Group and University of Valencia students
Adrián Merino and Balbina Zanon from the University of Valencia completed a successful internship at Connected Data Group, where they investigated how AI platform MindsDB and Logical Data Management platform Denodo can be used within the Connected Data Cloud platform. They developed a predictive model for healthcare and asset management, created a practical application guide, and discovered the great value of AI and data virtualization for real-time analytics. Afterwards, Adrián joined Connected Data Group as a junior data engineer, while Balbina is considering pursuing a Master’s degree in AI.

You may have studied BI and analytics for four years at the University of Valencia, but that does not mean that you can also apply all that theoretical knowledge in practice. Adrián Merino and Balbina Zanon therefore concluded their bachelor’s degree with a three-month internship at Connected Data Group, an Open Line company. During that time, they investigated how to best use the innovative data management and AI solutions Denodo and MindsDB, as part of the Connected Data Cloud as a Service, in practice. It became a win-win for the students and for Connected Data Group.
They both obtained a bachelor’s degree in BI and analytics, a course that combines economics and analytics. Adrián was attracted to the technological side of the field when he started his studies, while Balbina thought that she found the economic approach most interesting. “But it soon became clear that I have much more affinity with the IT and data side of the field,” she says.
Getting a taste of practice
When both finished their education in the summer of 2024, they asked themselves: should we look for a job right away, or should we first develop ourselves a bit further? Balbina: “I was quite unsure about what our theoretical knowledge would be worth in practice. That’s why I decided to do an internship first.”
Adrián was unsure whether he wanted to continue with a master’s degree right away, or first get a taste of practice. “Because you gain so much new knowledge in four years. But as Balbina already indicates, you don’t really know what you can do with that knowledge in practice.”
A fellow student had done an internship at Open Line Valencia and introduced the duo there. They passed the ball to Connected Data Group in the Netherlands. A golden pass, as it soon turned out. Adrián: “When we met Erik Fransen, managing director of Connected Data Group and specializing in data management and AI, we were immediately convinced that we could learn a lot from him. But we were also quite overwhelmed by the depth of the projects they do and the enormous amount of knowledge and experience they have.”
Research into Denodo Data Virtualization and MindsDB AI application possibilities
Connected Data Group had a clear assignment for the duo: find out how we can use MindDB’s AI technology and the Denodo virtualization tool and what that entails. “We knew straight away: we can learn a lot from this project, but it won’t be easy. We’ll really have to put all our energy into it,” says Balbina.
The internship lasted three months. Quite a short time to get to know two new products and learn to work with them. “But Erik, co-founder and CTO of Connected Data Group Antoine and the rest of the team immediately gave us a lot of confidence. We worked from Valencia and had contact remotely. Yet it didn’t feel distant, their door was always open, figuratively speaking, and it was very easy to ask questions.”
Developing a AI-driven predictive model
They got to work on a fictitious case: predicting the course of the disease of a certain group of patients in the hospital based on various types of static and streaming data. The data was freely available for testing. Adrián: “We wanted to be able to make predictions about the length of stay and the chance of complications based on machine learning and large language models, and we wanted to visualize the results in PowerBI afterwards. MindsDB is a very powerful low-code AI-platform that you can use to develop predictive models yourself. There is a community on Slack that has a lot of knowledge about the product. That was a main source of knowledge.”
Writing a manual
Before the duo could start making analyses, the datasets had to be unlocked and integrated in Denodo and MindsDB. And that was quite challenges as the tools where new to us, says Balbina. The integration with the visualization solution PowerBI is also not exactly a piece of cake. “You can’t find a ready-made manual on the internet anywhere on how to do that. We looked into everything ourselves and then wrote that manual for Connected Data Group, so that colleagues who want to use this technology can make more progress.”
It soon became clear that the combination of Denodo, for virtual data integration, and MindsDB, for powerful yet low-code based AI, has a lot of added value for Connected Data Group and its customers. “Especially because this solution is very good at analyzing real-time data. In addition, there is built-in AI that searches for connections in data itself. The larger the datasets and the more varied the data, the more difficult it is for people to find patterns and deviations in them. AI is much better at that. With MindsDB you can really search for these connections in a targeted manner,” says Adrián.
The model in practice in a real life case
Although Adrián and Balbina based their research on fictitious data, the model they developed could be applied fairly quickly to a real case at an actual water board customer, which was very well received. “That was not about patients and predicting needed care, but about water board assets like wastewater treatment installations that needed maintenance. However, it turns out that the underlying method for arriving at relevant predictions is very similar. Instead of predicting which patients are most likely to have a complication, you predict which assets are most likely to malfunction and which intervention will then yield the greatest chance of success. The data you use is different, but the basic model was also applicable in this situation.”
Future
After her internship, Balbina went to Seoul for a Korean language course and is thinking about a master’s degree. Adrián joined Connected Data Group as a junior data engineer and wants to combine his job with a master’s degree at the University of Valencia. Both are grateful for the opportunity that Connected Data Group offered them. Balbina: “I now know much better what my theoretical knowledge is worth in practice and how I can apply it. That has stimulated me to continue studying.”
Adrián discovered that the combination of working and studying works best for him. “I like that I can apply new knowledge directly in practice.” He thinks it’s great that he now has that opportunity as a member of the Connected Data Group team. “I work remotely from the Open Line Valencia offices, but I am really part of the team in the Netherlands and come over every month to work onsite with co-workers and customers. I think it’s great that I can continue my studies here, but at the same time gain relevant practical experience. It’s something I would recommend to every student.”