Postdoc position @ Inria Rennes - Self-Tuning Algorithms for Hyperparameter-Free Optimization

[Paul Viallard]

Dear all,

We are advertising a fully funded 2-year postdoctoral position in the MALT team at Inria Rennes, supervised by Paul Viallard and Romaric Gaudel. Feel free to share with anyone who might be interested.

The position focuses on the design and analysis of optimization algorithms that automatically adapt their hyperparameters during training. The goal is to eliminate manual tuning by embedding online learning mechanisms, particularly from adversarial bandit theory, into stochastic gradient-based methods. This involves theoretical development (e.g., regret bounds, convergence guarantees) and empirical validation on real and synthetic tasks, with potential integration into widely used ML libraries.

Application deadline: 2025-08-31
Apply here: https://jobs.inria.fr/public/classic/en/offres/2025-08984

Feel free to contact us if you have any questions: paul.v...@inria.fr and romaric...@irisa.fr