Berlin, Berlin, Germany
• Planning & implementation of all relevant data structures for a successful calculation of CO2 emissions of various B2B customers. • Growing the data team (analytics, science & engineering). • Automation of calculation standards via machine learning methods (Feature Engineering, Imputation, etc.).
• Management of a team of four data scientists working in different verticals of the company (hardware, operations, fintech & consumer). • Identification of potential machine learning or data science related synergies between verticals and their prioritization. • Establishment of proper guidelines for data science related work (model documentation, review, etc.). • Iterative development and testing of modular deployment optimization algorithms as well as putting them into production (e.g. demand forecasting). • End-to-end planning and execution of multi-stakeholder analyses focusing on ground operations efficiency.
Knuper further processes anonymized mobile phone metadata to develop smart solutions for official statistics. We call it "Upcycling Data". Specific operations: • Development of an open-source python module based on Spark which extracts privacy-preserving statistical features from raw mobile phone metadata (CDR). • Creation of a privacy-preserving Convolutional Neural Net (CNN) for Federated Learning using Secure Multi-Party Computation (SMPC) • Setup and management of various cluster environments (AWS, HDFS). • Developing and maintaining several statistical models for the predic- tion of socio-demographic indicators at scale. • Multi-stakeholder project work in different African countries.