Gilles Backhus

Machine Learning Engineer
KONUX

Neural Networks in Sensors

At KONUX we develop sensor systems designed to operate in tough industrial environments. From a sensor's ground measurement principle all the way to long-term machine State of Health predictions, any signal processing task at KONUX is done using machine learning because of its numerous benefits such as cost, optimism and modularity. Our recent line of products have benefited in particular from using forward and recurrent neural networks because of they combine a gating-capable on-off philosophy along with interpolating high-precision capabilities which works particularly well in multi-sensor-multi-range sensor systems. Neural networks enabled us to solve old, hard sensor problems "over night".

From student to teacher. This non-technical statement is what drives me. I want to help industries to shift from old ways of R&D to the new era of machine learning with all its benefits. At KONUX we develop intelligent sensor systems, all the way from hardware to analytics, where I am responsible for machine learning related developments. I see myself rather as a teacher to learning machines than a student working out solutions using my own brain. I received a M.Sc. in EE at TUM as well as an Hrs. Degree in Technology Mgt. at the CDTM, both in Munich.

Buttontwitter Buttonlinkedin

Connect

RE•WORK on Twitter

Unavaliable

Be Sociable

  • Twitter
  • Facebook
  • Linkedin
  • Youtube
  • Flickr
  • Lanyrd
  • Instagram
  • Google plus
  • Medium