David Neumann, M.Sc.

Relevant Work Experience

• June 2019 – Present

  Research Associate at Fraunhofer HHI

  Berlin, Germany
  • Contributed to research efforts in bridging the gap between local and global explainable artificial intelligence (XAI) methods.
  • Helped in the development of novel approaches for neural network compression.
  • Aided in the development of the first international standard for neural network compression (ISO/IEC 15938-17:2022) that was spearheaded by the Fraunhofer HHI in collaboration with multiple international teams from research and industry, which resulted in international patents.
  • Worked on federated learning projects, focusing on reducing communication overhead using neural network compression techniques.
  • Developed methods for training recommender systems using federated learning, enabling the secure and private on-device processing of user data.
  • Currently working towards my dissertation on the topic of clustered federated recommender systems, which will be the culmination of my research efforts into federated learning, neural network compression, and recommender systems.
• April 2016 – May 2019

  Research Assistant at Fraunhofer HHI

  Berlin, Germany
  • Conducted independent research into deep learning-based motion interpolation and developed a novel method for inter-frame prediction using deep convolutional neural networks (CNNs).
  • The research focused on exploring the potential of CNNs in achieving high-quality motion interpolation, with an emphasis on researching and evaluating different architectures and techniques.
  • Research resulted in a software prototype that utilized Python and the TensorFlow framework to turn low-frame-rate videos into high-frame-rate videos, demonstrating the feasibility of using CNNs for motion interpolation.
  • Authored a master's thesis based on my research, which detailed the deep learning-based motion interpolation method and compared it to traditional motion interpolation techniques.
• September 2012 – Present

  Co-Founder and Software Engineer at Krischbin, Neumann, Rögner & Voigt GbR

  Berlin, Germany
  • As co-founder, I contributed to growing the company, building client relationships, the planning and execution of projects, and business accounting.
  • Worked on a range of different projects, including:
  • Video-on-demand streaming platform where I worked on the creation of a scalable architecture for streaming high-quality video content to a large user base.
  • Business intelligence report platform for railway sector, where users could create building blocks of information, making them reusable in various business intelligence reports, and offering them in an online marketplace.
  • Device management software that helped businesses to keep an inventory of devices and optimize their utilization, thus reducing costs.
  • Car leasing platform, where individuals and companies could resell their car leasing contracts to other interested parties.
  • Web platform for local electricity provider with customer onboarding, management, and billing.
  • Currently, I retain partial ownership in the company, but am not involved in any projects, allowing me to focus on my research position at the Fraunhofer HHI.
• July 2015 – December 2016

  Co-Founder and Software Engineer at Taask UG (haftungsbeschränkt)

  Hamburg, Germany
  • Co-founded Taask, a platform that connects neighbors for household services, and was one of its lead software developers until its closure.
  • Collaborated with co-founders to design and develop the platform, including:
    • Development of a scalable backend infrastructure.
    • Designing and implementing a user-friendly UI/UX for the web frontend.
    • Integration of payment providers and other necessary tools to enable seamless transactions.
  • Worked on guerilla marketing campaigns to attract users and build community awareness.
  • Developed valuable skills and insights through this experience, including the importance of bootstrapping a startup, market validation, and resource allocation.
• January 2012 – March 2016

  Research Assistant at Hasso Plattner Institute for Digital Engineering gGmbH

  Potsdam, Germany
  • Designed and developed a software solution for recording university lectures, enabling seamless integration with tele-teaching and online course platforms.
  • Collaborated on the creation of a portable recording box, allowing lecturers to easily capture high-quality video and audio content.
  • Implemented features for synchronizing presentation slides with audio and video recordings, improving the overall learning experience.
• January 2011 – December 2011

  Student Employee at th data GmbH

  Berlin, Germany
  • Contributed to the development of software solutions for the supply chain management in the automotive industry.
  • Designed and implemented a custom application for utilizing colposcopes, which are medical cameras used in a medical diagnostic procedure called colposcopy.