Uwe Naumann

RWTH Aachen University & NAG Algorithmic Differentiation Technical Consultant

Uwe Naumann has been a professor of Computer Science at RWTH Aachen University, Germany, since 2004 and a member of NAG since 2010. He received a PhD in 
Applied Mathematics from the Technical University Dresden, Germany, in 1999. Post-doctoral appointments brought him to INRIA Sophia-Antipolis, France (1999-2000), the University of Hertfordshire, UK (2000-2002) and Argonne National Laboratory, USA (2002-2004). His research and software development is largely inspired by Algorithmic Differentiation (AD). His areas of interest range from numerical methods and adjoint parameter sensitivity analysis via simulation software engineering including computer program analysis and transformation to combinatorial scientific computing. 

NAG's collaboration with Uwe dates back to 2000 when work on the "Differentiation-Enabled Compiler Technology" project started. Funded by EPSRC, the project was extended three times during 2002 to 2014. It resulted in various AD-enabled versions of the NAG Fortran Compiler as well as a range of successful applications including TELEMAC-MASCARET (opentelemac.org) and ICON-O (mpimet.mpg.de). This academic collaboration extended more recently to the EU projects Scorpio and Fortissimo. Proposals for two potential future EU projects (ELLIS, CONFIRM) are currently under review. Over the years several peer-reviewed publications were produced jointly with former and present NAG personnel (Steve Hague, Malcolm Cohen, Jacques du Toit and Viktor Mosenkis, a former PhD student of Uwe's).

In 2017 Uwe spent a one-semester sabbatical at NAG's Oxford office working on Adjoint Code Design Patterns and resulting in a corresponding publication in ACM TOMS. During the same period he taught a course on AD as part of Oxford University's Master program in ''Mathematical Modeling and Simulation.''    

As a technical consultant for NAG Uwe plays a leading role in the delivery of Algorithmic Differentiation Solutions to NAG customers, including an already substantial and growing number of tier-1 investment banks. He supervises an Aachen-based team of NAG collaborators (Klaus Leppkes, Johannes Lotz. Simon Märtens, Patrick Peltzer, Jan Riehme) focusing on the development of the AD software tool set dco (dco/c++, dco/map, dco/fortran) and making contributions to the NAG AD Library

Learn more about NAG Algorithmic Differentiation Solutions.