INSIDE EOSC 05 – Julien Homo
Andrew Dubber
Hi, and welcome to Inside EOSC, a podcast all about the inner workings of the European Open Science Cloud. I’m Andrew Dubber. I’m a professor in innovation, senior researcher at the Industry Commons Foundation, and I’m a consortium member of an EOSC project called LUMEN.
And every month I introduce, with no apologies for the terrible pun, luminaries, people who are central to the LUMEN project, and of course, the wider EOSC ecosystem. Today’s guest is CEO and data architect of FoxCub and technical coordinator for the LUMEN project, Julien Homo. Thank you so much for joining us.
Julien Homo
Thank you for the invitation.
Andrew Dubber
Fantastic. So I’m going to quickly ask about FoxCub.
I detect a slight French accent. Am I correct?
Julien Homo
Yeah, I’m French.
Andrew Dubber
Okay. And FoxCub is in Paris. Exactly. What is FoxCub?
Julien Homo
So FoxCub is a SME dedicated to support the public institution and organisation to work with semantic technologies and knowledge graph at scale. So the main idea behind this company is to share with this public organisation our experience based on all the work we already done for some private companies, some innovative solution we have deployed at scale in some other context, and with the idea to provide some support, sometimes very expensive when you are in our private domain, to help the public organisation to benefit of this kind of new technologies. So this is our idea. We want to help the public organisation to grow through this area from a technical point of view.
Andrew Dubber
Right. And is there a lot of commercial work for knowledge graphs for public organisations?
Julien Homo
Maybe, yeah. We want to reuse all the knowledge we collected through the different private experiences, the good practises, everything which according to us works well in the private domains to help this public organisation to progress, to work on their business model, for example, to be more sustainable, but also to be more competitive in this European context, for example, but also for national institutions.
So it’s our strategy. But of course, we are at the core of our company. We want to work around the semantic artefact technologies, the knowledge graph, of course, but also through the different AI technologies, because we are mainly a team of data architects with a strong background for these technologies.
So that’s why we are some experts to explore this new innovative area with these institutions.
Andrew Dubber
Right. So is most of the knowledge in this space primarily in the private sector?
Julien Homo
Yeah, not only, because we have diverse experiences between the teams behind FoxCub. I mean, some of us have a strong experience from private companies, but also we have researchers that have a background from public institutions to help us to cross the two domains, I mean, public and private, to coordinate our strategy in order to be sure that what we want to propose and to share is relevant for, of course, the public sector.
Andrew Dubber
You’re the CEO of a commercial SME, but you’re involved in a European-funded project. I have to ask why. What’s in it for you?
Because European-funded projects are not the world’s greatest get-rich-quick scheme.
Julien Homo
It’s an excellent question. And of course, we really want to help and to support this kind of organisation because we believe that this is the right way to do. We have noticed some gap between the innovations, between private sector and public domain.
That’s why we believe that we can do something to help this institution. But of course, it’s not easy for a private company to work in this condition because initially, the national or European projects are dedicated to build some projects without any benefit. It’s just when we collect some fundings, it’s only for personal costs or for travel costs, and that’s all.
Andrew Dubber
You have to prove that you’ve spent the money.
You can’t just keep the money.
Julien Homo
Exactly, but also we cannot invest. And this is something we have to deal with because it’s not really the best business model for SME to just spend the money and that’s all. No, we have to prepare the future.
So it’s more difficult to be fully based on public projects without any other partner from the private domain. We are trying to do this, but it’s more difficult for us to grow because if we want to hire, it’s a risk because we have to deal with all the constraints from the public project.
Andrew Dubber
I asked Suzanne Dumouchel, who’s the coordinator of the LUMEN project, when I interviewed her for the very first of these podcasts, what it was about LUMEN, because ultimately LUMEN is about building a software system that is transferable across domains, and we’ll go into that in some detail, but it’s 7 million euros of public money. It has 21 partners. What’s the hard problem that makes this not just a software commission?
Julien Homo
You mean why we need 7 million euros to produce?
Andrew Dubber
That, but also why we need 21 different organisations to contribute to what is essentially the delivery of what looks like it might just be a search platform.
Julien Homo
It’s another very relevant question, especially because from our point of view, with such an amount of money, like 7 million euros, in the private sector, we can, with a team with 5 members, we can produce a very important software at scale. But I guess a very important part of the LUMEN project, but more generally for the European project, we need to explore all the possibilities to allow the organisation to reuse what we have to produce. We have to explain, we have to demonstrate, we have to collaborate if we want to succeed in our approach, because it’s not just develop a software and sell it, it’s also allow all the organisation to adopt it, and also to expand it from a sustainable point of view without any additional cost.
So that’s why we need other partners. We have to also collect from them important feedback, because this is another way to work inside the organisations themselves, from the public sector. So it’s very important to understand how they work, and to work together, to collaborate together, if we want to develop something useful for the public organisation.
That’s why we need some partners, but you’re right, I’m sure that we can identify some optimisation with this amount of money to develop, I suppose, more efficiently the different features involved in the project. So yeah, it’s an important question, but it’s not the same purpose between what we could be done with this money in the private sector in comparison of the European projects.
Andrew Dubber
You could make something amazing for that money with a small team, but it wouldn’t do what this is trying to do. Is that what you’re saying?
Julien Homo
Yes, maybe, but at the end, we have to sell it, and it’s another problem, you know. So, okay, we can do a lot of things, but at the end, we have to provide something useful, and that’s why I believe we need to work. This is, for me, a very good approach through the European project, to work together with teams from different countries, with different approaches, to work, to collaborate, and I’m really convinced that, thanks to this kind of collaboration, we can build something useful for the organisations.
And you cannot solve it, or you cannot, I’m not really convinced that, even if we have 10 or 20 million euros to produce something, if we are just a small team, of course, we will do something very fun, amazing, but I’m not really sure that we can sell it easily, or if it could be finally useful. And then, of course, there is an important point, we want to be useful for the communities, for the researchers, the students, we want to improve the way how we can discover the knowledge. So, we are talking about money, but it’s also a philosophy and something we want to share in terms of knowledge and also capabilities for everyone.
Andrew Dubber
You’re coming at this as technical coordinator for the LUMEN project, but you’re coming at it from industry, from the commercial sector, and you’re working with a lot of academic researchers, people in public institutions, research organisations. Is there a culture clash? Is there a trust issue?
Do people kind of treat you differently from that domain? You’re nodding it.
Julien Homo
Yeah, of course. It’s also very interesting because, for example, when I was working for EDF, which is an electric company in France, the main important one, when we have to work together, I mean with a team of 80 partners, the main important thing is the methodology. So we have to use some framework, some agile framework at scale to be all coordinated.
So every team, I mean, for example, with eight teams of 10 people, we have synchronised agile sprints together at scale with a lot of rituals, with a lot of meetings every month, every morning, you know, to be all synchronised. So, I mean, this is something very huge. And if I have to compare, this is something totally different when you are in a consortium like LUMEN, because this is not the same temporality, the same way to work.
But at the end, for example, through all the meetings and the steps and the cycles, the procedures we have included in this methodology at scale, we have also introduced a huge latency because all of the team were mainly in some meetings to discuss about what we have to do instead of produce. Doing it. Doing it.
Yeah, you’re right. So it’s another culture with some other constraints. In this context with LUMEN, obviously we have to be sure that every partner is committed, but it’s another way to do it.
And it’s very interesting to compare the two approaches.
Andrew Dubber
I’m interested because what you just said reminded me of, I’m based in Sweden and in Sweden there’s this, what I consider to be a strange thing where they have meetings and then they have decision meetings. And those are two totally different types of things. And I thought, what’s a meeting that isn’t a decision meeting?
But apparently that is a thing. And it feels like that disconnect between what I’m used to, which is very kind of action orientated, right? You go in, somebody makes a decision, everybody goes off and works from that.
It makes me wonder about the fact that this is a European project. And not only does the commercial world and the public sector world have different kinds of ways of doing things, but also different countries and different, obviously even within Europe have these different cultural approaches to how things are managed. And so I guess my question is, as a technical coordinator, how much of what you’re doing is managing the technical side of things?
How much of it is managing people?
Julien Homo
It’s a really interesting question because indeed you have to identify the core team, I mean the expert through the different countries, and then you have to organise the team. So I guess the main important part in this kind of project is to be able to split properly the different features you have to develop. So instead of having a kind of horizontal approach, you have to be more vertical, I will explain.
It’s simpler to organise the teams through a dedicated part of the component we have to develop through an European project, because of course we have to deal with the distance, the culture sometimes, this kind of thing. So maybe it’s easier to share the responsibilities of an important part of some component. Instead of in some private companies, we can split the same component between different teams, because we are, I suppose, closer and we have not the same timeline.
But behind we have the same rules because we are in the same company. So it’s another kind of organisation, but at the end we have some solutions, but we have to be flexible. This is what I have in mind.
And then I’m really convinced that this is the main difference between these two domains.
Andrew Dubber
Okay. What’s LUMEN for, in your understanding? I mean, I’m part of LUMEN.
I’ve heard LUMEN described to me from different people, but from somebody who is charged with building the technology that LUMEN essentially represents, how will you know that you’ve done it correctly? What’s it for? What’s the outcome?
Julien Homo
Okay. In LUMEN, we are building a framework, which is the data mesh. I will explain today the scientific communities, for example, molecular dynamics, SSH, Earth systems, or for example, mathematics in France, but also in Europe, have already developed very relevant tools, services.
They are already sharing data, also datasets. So, okay. Each community has already developed very relevant products or services for their own community.
But with LUMEN, what we really want to address is the interdisciplinary use cases, how we can cross some usage between these communities. And because each community has its own specificities, its own culture also, its own way to work. So, the main challenge for LUMEN is how we can address, how we can build cross usages between these communities, bridges between them, but not also usages, how we can reuse some very interesting work, great work already performed, already shared, already released by a community to help another one to grow, to use this platform for their own purposes.
So, with LUMEN, we are building this federation with a federated infrastructure to help to build new cross usages, but also to share these existing services between the scientific domains. And also to reuse some part of this component to instantiate them in the different communities. So, from a technical point of view, we are trying to identify the most relevant component of each community to help the other one to adopt them.
And it is the case for GoTriple, for example, from SSH. We want to allow some communities to reuse this discovery platform for their own purposes, but also to build some bridges between this community, thanks to this kind of platform, but also thanks to other tools.
Andrew Dubber
Right. You said data mesh at the start of your answer to that question. And I’m not sure by the end of the answer to that question, I understood what you meant by that.
When you say data mesh, what exactly do you mean?
Julien Homo
The data mesh is the paradigms behind what I have just explained. So, instead of creating a new platform, another centralised platform where we have to put all the data together in the same space, in LUMEN, we want to be federated. I mean, we want to share data thanks to a kind of framework to be able to enhance the interoperability between the communities thanks to new rules.
And this model, this paradigm is called, in our context, a data mesh. A data mesh is a kind of model of federation to share data, to share product together, already deployed in some private sector, for example, with PayPal, to share their data product more efficiently and to improve this interoperability and the discoverability of their services. So, it’s just a federation model to be instantiated also from a technical point of view, but also from a governance perspective.
And yes, this is just a framework to share data and to improve the interdisciplinarity.
Andrew Dubber
It’s the rules that make the mesh?
Julien Homo
The data mesh is a framework with some rules, guidelines, concepts.
This is a kind of toolbox, and we can use some components from this model, from this framework, to allow the community to be more efficient, to share data, services, tools together. But we have also some other models to share these resources from a federated perspective. We have also, in the European context, the concept of data spaces with some differences.
But at the end, this is the main idea. We want to share more things together without creating a new platform for everyone or something in the middle, because we know that it doesn’t work. So, yeah, this is the idea behind the data mesh.
Andrew Dubber
Right. So, what’s the relationship between the data mesh and knowledge graphs?
Julien Homo
Okay, so the data mesh is just a framework to be more interoperable with a kind of toolbox.
Andrew Dubber
That’s about fair data, primarily.
Julien Homo
Yeah, then, the knowledge graph is just a space where we want to put some entities and their relationships to explore knowledge.
So, it could be a technical solution. For example, a database could be a knowledge graph. Also, it could be a platform.
For example, when we are talking about Wikidata, it’s a knowledge graph. So, we have different levels of what is a knowledge graph. But finally, this is, at the end, the same idea.
It’s just a space, a technical space or a functional space, where we want to explore data across a graph structure, I mean, entities or nodes and relationships between these nodes. So, yeah, it’s not at all the same concept that a data mesh, which is a framework or a model for collaboration.
Andrew Dubber
Here’s where it starts to get confusing for me, which is that knowledge graphs are based on the idea of the triple from ontologies. So, you have this thing called a triple, which is the subject, the object, the predicate, the integer. I get that.
I’ve had it explained to me very well. And LUMEN is based on, well, it builds on a project that came before, the GoTriple project. Am I right in saying that GoTriple didn’t actually use triples?
Julien Homo
No.
Andrew Dubber
Why is that? And is this what we’re trying to fix now?
Julien Homo
Yeah, you’re right. In fact, we have different ways to share data and to consume this data. And GoTriple, initially, the main purpose of this platform, this solution, is to expose this data through an interface, which is the GoTriple interface.
Andrew Dubber
When you say expose the data, sorry, basically, it’s a search engine for social science, humanities research.
Julien Homo
Exactly.This is a search engine. At the end, this is a search engine. But also, you can access the data through APIs.
So we have different ways to explore the data exposed by the GoTriple. But yeah, mainly, this is the search engine. But it’s important to take it into account because we could imagine, also, convert, translate this data into a graph in order to explore the content of GoTriple through another structure, which could be a graph.
But this is the difference. In GoTriple, we harvest data from data sources in the SSH domain. In order to index this data, I mean to convert the data in a structure compatible with the search engine.
Because when you are looking for something, you need a dedicated database to retrieve the information with some prerequisites, which are we have to be quick, to be fast, to be efficient, and to be relevant. So to do it, we have some dedicated database, and this is the search engine. But if we want, for example, we can explore the data in different ways, and it could be through a graph, but for other use cases.
And also, if we want to be compliant with some other mechanism, for example, if we want to connect the data of GoTriple with another knowledge graph in the European context, for example, maybe we have to translate to convert the data into triple. Because the triple approach is used to connect the knowledge graph together, mainly, of course, just to, it’s a very quick summary, but this is the idea behind. So if we want to connect the data from GoTriple with another discovery platform, a way to do it is to use the triple approach, I mean, the graph approach to do it.
Andrew Dubber
But just to be clear, when you say it’s a discovery platform, a researcher can use it, I’m talking about GoTriple at this point, a researcher can use it to find social sciences and humanities publications and papers and so on. It’s actually one of the, I’m not going to say rare, but one of the, it’s not common that an outcome of a European project of that nature is that useful when it’s finished. GoTriple is really good.
And I say that as somebody who researches in the social sciences and humanities. And the idea is that how good it is, the function of LUMEN is to expand that across other domains and actually to bring it to mathematics and molecular dynamics and presumably all of the other sciences in the world of EOSC ultimately, although not within the scope of this particular project.
Julien Homo
Not exactly, because, OK, you’re right, for example, when a domain doesn’t have a discovery platform or a robust one, the idea behind LUMEN is to allow these communities to get this platform to improve their discoverability across the EOSC ecosystem, obviously. But sometimes some scientific domains are more mature, you know, so in our context of LUMEN, mathematics has already its own discovery platform, which is the ZbMath Open . So we don’t want to impose anything through the LUMEN project.
And this is another principle of the data mesh. The idea is to be flexible and to allow them to reuse some part of the shared works or services, but they just have to reuse what they need to improve their existing services. So this is the idea behind the federation.
And just to explain that, for example, for mathematics, they are interested about what we are currently developing through the chatbot in the White Label platform, so GoTriple. But also they want to maybe reuse the annotation or classification services already deployed in GoTriple. So we can also share some part of the platform, some features.
But this is something important I would like is we don’t want to use the GoTriple model for every domain, not at all. It’s just how we can be more flexible and be more efficient together and to reuse what it’s worked through the European content. And maybe I could add something.
This is also why when we talk about LUMEN, we notice that this is an important trend to be federated. This is exactly what we want to build through EOSC, thanks to the different nodes. We want to federate some nodes to share the knowledge, to share the services.
And with LUMEN, we want to do it based on a dedicated model for the communities. But the idea behind is the same. That’s why we just want to share what works, but we won’t impose some, you know.
Andrew Dubber
That sounds like a really hard problem given particularly, okay. So you use the phrase white label, which in my mind is we have a solution. We’re not going to put our branding on it, but basically you can just apply it in your domain and it’ll work.
And that’s not what you’re saying. What you’re saying, like mathematics, you’ve got your own thing going on. Maybe some of this is useful.Maybe some of it’s not useful, but let’s adapt. And, you know, molecular dynamics, maybe you’ve got some things in place, but here is the thing. And so I’m wondering the extent to which a white label thing can be a white label solution, but also how much of a platform that is developed in one domain, like social sciences and humanities, for instance, is even applicable in something that seems, at least on the face of it, to be radically different in terms of how it works, what its terminology is, you know, how people use it, like say mathematics, for instance, which, I mean, it’s even written with different symbols and characters, you know?
So how transferable and how white labelable, if that’s a word, is LUMEN?
Julien Homo
Yeah. In fact, of course, there are some limits in this exercise, but the main idea behind this white label platform is we share some common concepts. So for each scientific domain, we have publications.
We have also data sets, semantic artefacts, I mean, vocabularies, ontologies, something like this, multimedia assets. Also, we have the concept of project, author profiles. So these are common concepts and we want to capitalise on it.
So this can be reused through the different discovery platforms across the different domains. But of course, we have to extend. Sometimes we have some domain-specific types.
For example, we have the molecules, the temperatures for the molecular dynamics domain. In Earth, we have geodata sets. And so we have to build something flexible and adaptable for each community.
But at the core, the workflows are the same. I mean, we have some common concepts and of course, we have to extend some of them for the purpose of each domain. But also, for example, I already mentioned the annotation service, which is something to help to tag the resources of the discovery platform with an external vocabulary.
This external vocabulary has to be provided by the community. So we have to work with them to understand which kind of tag we want to apply on their documents. But it’s also the case with the classification service.
We have to identify how we want to classify the resources in the context of this domain. So, yeah, we have a lot of things we cannot generalise as a white-label thing, but we have a lot of things we can put in the common space.
Andrew Dubber
There’s another layer of complexity to this in the idea, and the word serendipity gets used a lot, the idea that discovery between domains can happen. If we essentially build everything on using the same railroad tracks, so that you can use some knowledge from mathematics and some knowledge from molecular dynamics to come up with kinds of research that would otherwise have been impossible. And that somehow LUMEN allows for that to happen.
So instead of simply building something for each domain, you’re building something that joins those domains as well. How hard is that as a problem, technically speaking?
Julien Homo
You’re right. So in order to solve it, for example, another component, another key component in this LUMEN project is the meta-search, its role is to be able to identify some cross-use cases between the domains. For example, how we can identify the impact of some biologic solution developed by some companies from a societal point of view, something like this, something between SSH, molecular dynamics, and maybe with the Earth system domain.
So, OK, maybe we have some publications related to this topic, I don’t know, and the meta-search, its role is to be able to identify this is a cluster of publications for this cross-domain use cases. OK, it’s a technical challenge to identify this kind of clusters, but to do it, we use the data mesh paradigm with some key concepts into it. So, for example, instead of having different kinds of resources, I already mentioned the publication, the data set, the semantic artefacts, we are talking about data products and data contracts.
So we are able to just identify the same metadata for the same community and the rules to share it together. Thanks to it, we can apply some automatic algorithms and also some automatic programmes to explore this knowledge easier, thanks to this kind of approach, I mean, a common data model shared by the communities in order to use the last innovative approaches to explore automatically this data. And to understand better how we can use it and to discover new things, because we don’t have to know what are the different kinds of resources exposed by the community, because this is inside the same model, you know, everything is encapsulated into a data product and we just have to explore all the data products to discover something.
Andrew Dubber
I’m wondering what’s interesting about this to you. And the reason I asked that question is, one is you explain all of the aspects of this really, really well. You’re really good at what you do in terms of how you manage the technical aspects of this project.
There’s clearly something that you’re passionate about in this. And I’m wondering whether it’s the solving of a technical problem or whether it’s the different domains that you get to find out about, or there’s something driving you. And I’m wondering where that comes from and what your background is.
I feel like you’ve come from somewhere to this and you’ve gone, that’s what I want to do. That’s how I want to be spending my time. Because you seem really interested in it.
And you seem to be having fun with it as well. Not just you’re good at it, but you’re enjoying it. What’s your background and what brought you to this?
And what’s the thing in this that you find really worth your time?
Julien Homo
OK, I don’t know if it’s easy to answer, but if I try, for me, what I like is how we can modelize the knowledge. So that’s why today I have a master’s degree in knowledge representation, modelization and in artificial intelligence. But in 2011, you know, we didn’t have any ChatGPT model or LLM approach at this time.
Andrew Dubber
We had a good 50 years of AI development, but we didn’t have what we have today.
Julien Homo
Exactly. It was another period for this AI part. But before, I was very interested in the ability to how we can represent high end things, how we can modelize everything.
And with our important conviction that, OK, we have different ways to do it. And obviously, we have to manage it. And I’m sure that we don’t have only one way to do it, of course.
But the challenge behind, I mean, how we can reconciliate the different ways to modelize everything. I mean, the knowledge. It’s something I like to explore.
Andrew Dubber
Because it feels like one of those projects that can never be finished. You know, how will we represent all of knowledge? And I think that there’s, you know, how the brain works, how knowledge is created, what knowledge is.
I mean, the fact that it’s not just a collection of facts or even relationships between facts. But do you get the sense that we’re in a time where the representation of knowledge is going through some sort of great leap forward? And, you know, I guess, do LLMs play into that in a particular way?
Julien Homo
OK, so thanks to the LLM, of course. For me, it’s an innovation, obviously, because we have some chat based interfaces, really user friendly. And it’s a way to facilitate the knowledge exploration.
So in that sense, for me, it’s an innovation, a very important one. But also, we need to understand better how we can improve these LLMs, these models. Because we have seen that, OK, we have some bias in this approach.
Sometimes we got some wrong answers. We have to, you know, clarify the sources used to generate this chat. So it’s now that we need other structures to help this model to be more relevant, to be more precise, to improve the trust.
So to do it, we need the graph. We are exploring some different approaches, but one of them is, OK, maybe the ontologies can help the LLM to work better. We have some other technical approach to do it.
The RAG, some other algorithm based on the graph solutions. But thanks to the LLM, we need to reconsider another approach, which are, for example, the ontologies to facilitate the operation or to improve the trust into it. So what I mean is, again, we have to reconciliate different kind of modelization to help the end user to get the knowledge, to explore and to learn more, you know.
So I guess we don’t have the best solution. It doesn’t exist. We have just to find the right balance between all the possibilities, the option to represent the knowledge and we have to support it.
So, yeah, this is a continuous improvement, you know. Yeah.
Andrew Dubber
Okay. I want to, I want to finish this by bringing it back around to where we started on the fact that you’re a commercial organisation, but you’re working in the domain of open science. It’s a European Open Science Cloud project.
What is the value of open science to a commercial organisation? Or to put it the other way around, what’s the value of commercial activities to open science? Either of those is fair game.
Julien Homo
Okay. So thanks to the open science and more generally, thanks to this kind of European and national project, we can try, we can experiment something. And this is something really important for us.
I mean, for the private companies, because this is where we can try something without important risk. You know, this is a very important approach for us to innovate and to create new things and then maybe to develop new solutions based on these outcomes and these results. So for me, it’s a kind of win-win approach for the project, for the open science.
We want to share what we have learned through our previous experience in the private sector. And then thanks to this kind of context, this collaboration with the other partner, we can innovate, find new ways to develop things for the open science and then reuse this result for some other commercial approaches, but in another context, you know. So that’s why for me, it’s a very interesting approach with another condition, which is we want to help the organisation to share more knowledge.
So it’s another part for the open science. We believe that it’s really important to share the data together. Of course, sometimes we have sensitive data.
We cannot share everything everywhere, but we can improve a lot of things thanks to this kind of open science and open data strategies.
Andrew Dubber
Okay. Well, speaking of being open, I’m going to finish this one question and answering is optional, but I’m curious, your company is called FoxCub. Yeah.
Why?
Julien Homo
Okay. Why? Because, first of all, there is a link with Firefox, which for us was a huge company in 2011, during this period, and we are not a fox.
So we just are FoxCubs, and we want to follow the path of this kind of important company, because for us, the Mozilla Corporation is a kind of open company to share knowledge through services and components. But that’s why we were…
Andrew Dubber
It’s a tribute.
Julien Homo
Yeah.
Andrew Dubber
Wow. Yeah, that’s really fascinating.
Julien Homo
Yes, this is one of the main reasons we decided to call it FoxCub, but yeah.
Andrew Dubber
That’s really lovely. Julien, thank you so much for your time. I really appreciate it.
And yeah, I look forward to another couple of years working with you on the LUMEN project. Thank you, Andrew.
That is Julien Homo, CEO of FoxCub and the technical coordinator for the LUMEN project, by which this very podcast, Inside EOSC, is brought to you.
LUMEN, of course, receives funding from the European Union’s Horizon Research and Innovation Programme. I’m Andrew Dubber, and I’m back next month with an interview with another EOSC insider. You can subscribe to Inside EOSC at absolutely no cost to you whatsoever, wherever you listen to podcasts.
You can also give it five stars, leave a review, post about it on social media, if you’re one of the people who still does that sort of thing, or pass it on to friends and work colleagues. Very grateful for any and all of these things. Thanks so much for listening.
We’ll talk soon.