Pilot Preventive Maintenance Machine Learning Internship

The PositionInternship for 6 Months Project: Machine Learning / automated analysis of production data and anomaly detection for predictive maintenance applicationsTask:Roche operates a so-called data warehouse in which all production data is collected and analyzed to a certain degree. The data is obtained from hundreds of test fixtures that use various measurement techniques to ensure that the produced instruments and submodules adhere to their specifications. Currently the webinterface of this data warehouse allows to display said data and the backend provides basic functionality that can automatically detect anomalies in the measurement results and trigger alarms. Following an alarm, a production engineer reviews the data and determines whether further action has to be taken or whether the alarm can be ignored. The tasks for this internship are:Understand the existing methodology and interface to the data warehouseExpand the existing algorithms such that predictions can be made when specification limits are going to be exceeded. Include this information in the automated alarmsEvaluate the possibility of distinction between anomalies that require maintenance activities and anomalies that hint towards problems with devices under testImplement a machine learning algorithm that takes the feedback of production engineers and improves the anomaly detectionSuggest improvements to data collection and analysis, identify gaps in the current implementationRequirementsAbility to program in one of the languages python, R, java or java scriptTechnical background familiar with measurement methods and processesKnowledge of statistical data analysis methodsFluent German (due to existing documentation) and EnglishPreferred:Pre-existing knowledge in the fields predictive/preventive maintenanceStudies in engineering or scienceAbility to work independently with little supervisionWhat you can expectModern infrastructure of highest ecological and ergonomic standardsFlexible working environment (in offices as well as home office)Flat hierarchies where input is appreciatedOpportunity to operate in a vast and diverse network, to realize projects together and get insights in various areasWorking in a fast responsive core team with globally operating stakeholders Your complete application includes the following documentsA Current CV and a Motivation LetterA certificate of enrollment (if you are currently studying). If you want to use this opportunity to do a Master’s or Diploma thesis, please highlight this in your application Due to regulations, non-EU/EFTA candidates must provide a certificate issued by their university stating that an Internship is mandatory for the studies, and be continuosly enrolled at the university