The Clausthal Continuous Optimization Group at Clausthal University of Technology (TU Clausthal)
invites applications for two 4‑year full-time postdoctoral positions (E13 TV‑L),
funded by the VolkswagenStiftung Momentum project “LiftHighUp – Training algorithms with Lifted High-Rank Updates”.
The project focuses on developing and analyzing novel lifted second-order methods for deep learning, with an emphasis on structure-exploiting algorithms, numerical linear algebra, and energy-efficient training. Responsibilities include independent research at the interface of continuous optimization and deep learning, publication and presentation of results, development of open-source software, supervision of student assistants, and active collaboration with interdisciplinary partners at TU Clausthal. The positions offer the opportunity for further academic qualification (habilitation).
Applicants should hold an excellent PhD in mathematics, computer science, or a related field, with a strong background in numerical optimization or deep learning, have a solid publication record, very good programming skills in Python or Julia, and strong communication skills in English. Experience with deep learning frameworks and/or GPU programming is highly desirable.
Application deadline: 15 May 2026
Contact: Prof. Dr. Andreas Potschka (andreas.potschka@tu-clausthal.de)
Posted on 2026-04-29 by Magdalena Sroczan