I am joining Inria Grenoble in 2024 as a researcher

In January 2024, I will be joining Inria Grenoble1 as an Associate Researcher (my home-brewed translation of Chargé de Recherche) in the POLARIS team and wanted to make a quick note about it after exchanging about several points and clearing out confusions from several people.2

What is Inria, and that position?

Inria is a fairly large institution in size and scope. Broadly speaking, research topics vary across domains in computer science and (mostly applied) mathematics. The institution also places an emphasis on technology transfer and impact beyond one’s research domain. Some projects you may have heard of that started at Inria include the OCaml programming language and the scikit-learn library.

The position

As a rough projection if you are in the German system, think of this position as a W{1/2} permanent position at a Max-Planck Institute (as a researcher without teaching duties). Unlike Max-Planck, Inria has centers within university campuses with research groups that are mostly (always?) composed of permanent and temporary researchers and staff from research institutes (Inria, CNRS) and surrounding universities / grandes écoles.

One can make a parallel between the career steps as a researcher at Inria and CNRS and the ones at a university:

Inria University
Chargé de Recherche (CR) / Associate Researcher Maître de Conférences (MCF) / Associate Professor
Directeur de Recherche (DR) / Senior Researcher Professeur des Universités (PU) / Full Professor

I translated DR as Senior Researcher and not Research Director, which would be the most obvious translation, because unlike what the title implies, Research Directors do not necessarily have management-heavy positions and conduct their researcher at a smaller scale. This is different from a principal investigator (PI) position involving having one’s own group (more on that in the next section).

One thing to note on the type of position is that this is a public servant position, with a salary range regulated at the national level. Being hired at a national level also means it can be easier for instance, to move to other Inria centers, which wouldn’t be the case with contracts tied to the local university.

Finally, Inria opened in the last years a second type of permanent Junior position, Inria Starting Faculty Positions to recruit at the same level as CR but with different hiring, working and promotion conditions.

The project-team organization

One specificity of Inria is its organization in project-teams (équipes-projet), a group of tenured researchers and faculty at different levels of seniority organized together towards a specific goal or set of research questions. Project-teams can vary in size, from about 4 to 15 tenured researchers and faculty (numbers from the top of my head and empirical evidence, there might be examples stretching this bound). Think of it as blending a static organization (research departments and divisions in other places) with the temporary nature of a research project. A project-team can be reconducted to continue in similar or extended research directions, but is not expected to be set in stone. Project-teams are created, terminated, merged, split as researchers within them feel the need to reorganize the way they work.

Another key aspects, highlighted from the beginning of the excellent Advice Booklet for Inria Applicants, is that they are not “research chairs” or “groups” from the German or US models that would be centered around one senior Principal Investigator hiring and being in charge for everyone according to their vision. Multiple members are permanent researchers, collaborate and hire their Ph.D. students, postdocs, students, or engineers, and choose to collaborate within the team in various ways.

How does one get that position?

This question could fill a whole series of blog posts and I will not delve into too much detail. As an interesting fact to understand, the hiring process is a national competition regulated by a standardized open call with a fixed number of places per Inria center. Most applicants provide a research proposal with a heavy emphasis on its integration within one of the team-projects. I’ll refer to the Guide for Inria Applicants again for a more complete overview.

A last point on the matter, I would consider that “just applying” to these positions when they open is nearly impossible: contact the teams beforehand, several months before the call for applications. I am highly indebted to the people from different groups who went out of their way to give me thorough reviews and critics on the application, the audition presentation and my research statements.

Coming steps in the near future: settling and building

The first obvious thing on my side will be to manage to move all the random items I accumulated in Berlin to Grenoble, and then move myself there and somehow settle. I am still “amazed” at the sole quantity of books that I collected here and already filled a huge bag.

More interestingly, I will continue working on my core research topics in methods for mixed-integer and convex constrained optimization.
I will start to look for motivated students to start working on various topics in computational optimization both on nonlinear and mixed-integer aspects. If you like a blend of theory, algorithms, and programming (Julia, C, Rust, C++ as primary targets) on cutting-edge software, feel free to reach out. I will try to put out a list of potential projects to provide a good representation of possible activities.

I will also continue to work with people at the Zuse Institute on the whole bag of current topics we have ongoing and regularly come to Berlin. If you want to talk in person, just ask in advance to be sure I’m around, knocking on my office door will probably not suffice anymore. We will for instance merrily continue to push the methods in convex mixed-integer optimization within the MATH+ project that started this year, continue working on algorithms within SCIP and first-order methods for constrained optimization.

On the prospective side, I am looking forward to starting discussions and then collaborations within the vibrant Grenoble ecosystem. From combinatorics to quantum information and proximal methods, topics adjacent and connected to mine are not lacking, both on methodological interfaces and application areas. Working on the research proposal for the application had the benefit of making me explore in depth the activities of the people in my project-team, the LIG laboratory, Inria and other Grenoble institutions, and build connections with my mid-term research interests.

Rethinking about this whole year (we are in October, I started preparing applications in December last year), I’m pretty hyped about what is to come, also blended with a much more latent relaxed mindset. Getting a permanent position that would let me prioritize on the activities I like and find important had been one of the major things “running in the background” for a while, and it lifted the curtain on the possible paths to choose from. Rethinking of the alternatives, I do think that the position-then-scaling approach of that system is more sustainable than the scaling-then-position (getting growing management, funding and teaching responsibilities before as a proof that one is worthy of a faculty position).

  1. The actual full name is Inria Center at the Université Grenoble Alpes. And the second geographical affiliation is the Grenoble Informatics Laboratory↩︎

  2. Banner source: Wikipedia. 2023. “Grenoble.” Wikimedia Foundation. Last modified July 1, 2023. ↩︎

Mathieu Besançon
Mathieu Besançon
Researcher in mathematical optimization

Mathematical optimization, scientific programming and related.