Technische Universität Berlin - Faculty IV - Institute of Telecommunication Systems / Media Technology
Research Assistant - salary grade 13 TV-L Berliner Hochschulen
Berliner Zentrum für Maschinelles Lernen
under the reserve that funds are granted - part-time employment may be possible
The goal of the project is the exploration of compression techniques for machine learning models. Your research will focus on the development of novel dimensionality reduction and compression algorithms for neural network architectures (convolutional, recurrent, autoencoders). Furthermore, you will investigate new ways of integrating invariances (e.g., translation, rotation) into the neural architectures.
Successfully completed university degree (Master, Diplom or equivalent) in computer science, engineering, physics or applied mathematics
Profound knowledge in machine learning, in particular neural networks, efficient deep learning, invariant representation, and it application fields (e.g., computer vision, communications, medical image analysis)
Practical experience with training, compressing and applying neural networks (ConvNets, Autoencoders, ResNets etc.) for the analysis of image and video data.
Solid programming skills, in particular experience with deep learning frameworkds (PyTorch, TensorFlow etc.) and Python libraries (scikit-image, MayaVi etc.)
Excellent communications skills in English
How to apply:
Please send your written application with the reference number and the usual documents to Technische Universität Berlin - Der Präsident - Fakultät IV, Institut für Telekommunikationssysteme, FG Medientechnik, Prof. Dr. Wiegand, Sekr. EN 16, Einsteinufer 17D, 10587 Berlin or by e-mail to firstname.lastname@example.org.
To ensure equal opportunities between women and men, applications by women with the required qualifications are explicitly desired.
Qualified individuals with disabilities will be favored. The TU Berlin values the diversity of its members and is committed to the goals of equal opportunities.
Please send copies only. Original documents will not be returned.