Companies that have to manage lots of assets (railroads, ports, highways, or other infrastructure related institutions) have to inspect and administer the configuration and condition manually, multiple times a year.

Asset management is a labour-intensive job, error prone and reactive to changes. Failing assets come as a surprise and wrong parts are being ordered. Therefore a platform is needed that is able to perform inspection and check configuration of assets automatically and to enable a proactive approach. For example: the platform ingests image and video data, along with meta data, of multiple assets and runs several AI components and Machine Learning models to solve various problems, like asset recognition, segmentation, condition classification, and configuration correction. To make a proactive approach successful, new models and data need to be pushed to production as fast and as secure as possible.

Maarten will describe his experience with the CD4ML approach (Continuous Delivery for Machine Learning), a technique that allows the platform to quickly expand with new AI components and Machine Learning models, from development into production in no-time.

Talk at Hamburg Data Science meetup : Please register at https://www.meetup.com/Hamburg-Data-Science-Meetup/events/279530699