Every year, SequeL is looking for:
- graduate students to complete their PhD
- post-docs
- R & D engineers
- full-time researchers (junior / senior)
General schedule:
- applications are due at the end of winter/beginning spring, each year
- acceptance in the early summer, each year
- positions are actually open in fall, each year
So, for the next year, get in touch no later than winter!
You may also consider the following possibilities:
- staying with us during your sabbatical
Positions in 2017
Researcher positions at CNRS
- Deadline is January 6th, 2017.Check the CNRS webiste
SequeL belongs to section 6. Some words of advice
If you wish to apply, get in touch with philippe -dot- preux -at- inria -dot- fr.
You apply through the CNRS laboratory, that is CRIStAL.
You apply through the CNRS laboratory, that is CRIStAL.
Researcher positions at the INRIA (CR, DR)
Deadline for application is February 17th, 2017.
The application form will be available on January 16th, 2017 on the Inria website.
If you wish to apply, get in touch with philippe -dot- preux -at- inria -dot- fr
Postdocs and PhDs
Postdoc positions
- ML & system identification, supervision: Daniil Ryabko.
PhD positions
- Mechanism Design for Multiple AIs, supervision: Christos Dimitrakakis.
- Towards antifragile software: perturbations of software driven by data, supervision: Ph. Preux and M. Monperrus. detailed proposal, in English
- Online learning on streaming graphs
Internships
For master students (or equivalent) only (sorry, no undergrad).
- You love software development, graph theory, and machine learning. Below are two great proposals. Each is meant to start asap, last 6 months. This is a collaboration between SequeL and a software eng. group (Spirals).
- Modeling the evolution of software. Advisor: Ph. Preux & M. Monperrus.
In short and in English: we model software with (call) graphs; we have a generative model of such call graphs; we want to study how the call graph of a software evolve along commits/versions. This can be seen as a time series problem in which the series is made of graphs, and the model is (currently: (good) propositions are welcome) an algorithm. - Randmisation strikes back for the sake of software security, advisors: M. Monperrus, W. Rudametkin, and Ph. Preux.
The idea is to perturb at random a software code to assess its robustness to bug, and to attacks. As this is a problem of sequential decision making under uncertainty, the selection of points to perturb in the code is to be made using bandit algorithm.
- Modeling the evolution of software. Advisor: Ph. Preux & M. Monperrus.