Ruben Glatt

My research aims at combining the successful approaches of Deep Learning, Reinforcement Learning and Transfer Learning to improve the generalization of knowledge through abstract representation and autonomous similarity determination for intelligent multi-task agents.

Developing a monitoring system that uses object recognition in thermographic images to analyze certain regions of a target electrical system. The goal of the project is to automate and facilitate certain aspects of maintenance in a hydroelectric power plant.

Administration and development of datacenter control and monitoring systems of several datacenters in Germany and the US. Supporting datacenter infrastructure optimization based on datacenter monitoring analysis, statistics and evaluation of new technologies.

Supervision of the first 1&1 datacenter co-location in the US in New York (2003-2004) and training and support of staff for the co-location (2004-2007). Support and training of staff in the initial phase of building the first own datacenter of the 1&1 group in the US in Kansas (2007). Organisation and execution of server and system relocations in and between different datacenters in Germany and the US. Collection of statistical data about hardware in datacenters and co-locations.

Handling of all datacenter issues in a small team including maintenance of server hardware and datacenter infrastructure equipment. Technical contact for server homing customers in Germany.

Living up to the fast growth and high flexibility of a very successful internet startup company. Helping out wherever necessary from preparing faire brochures over billing to customer follow up.

My research aims at combining the successful approaches of Deep Learning, Reinforcement Learning, and Transfer Learning to improve the generalization of knowledge through abstract representation and autonomous similarity determination for intelligent multi-task agents.

Set up of a gesture recognition system based on a deep learning architecture utilizing Restricted Boltzmann machines and show how it performs when trained with changing multimodal input data on an Italian sign language dataset.

Development of small autonomous robotics platform (layout, sensors). Development of photovoltaics system (database integration last 20 years, continuous live updates+visualization) [http://www.pkm.kit.edu/english/3427.php]. Analysis of business and maturity models for technical service providers.