Tobias Genz

- Research, Concept and Implementation of a mathematical optimization service (Python, combinatoric optimization) as basis for a new product - Contributing for setting strategic direction for new product - Participation and leading of customer talks for product developement - Product demos to win new customers - Coordination of customer requirements together with development team

- actionable reporting and data analysis to support decision making across the company - cloud environment aws (aurora, postgres/mysql) using tableau, excel and jupyter notebooks - design and implementation of data pipelines - etl + smaller services/tools in python using aws stack (lambda, cdk,...) - setting best practices, strategy and vision for the data team - foster alignment with product+engineering teams

- Development of Computer Vision Algorithms used as Microservices with Python (opencv, tesseract, flask, tensorflow, keras, pyav/ffmpeg) - Micro Service and shared library API design - Unit and Integration Tests - Code Refactoring - Instance Segmentation of Holograms on ID Cards (Masterthesis)

- Text mining project (PoS tagging, word embeddings, topic models, sentiment analysis) - Image recognition project (CNNs, tensorflow, keras, object classification and detection) - Occasional teaching of Msc. courses: "learning from data", "probabilistic modelling"

(part time as working student from 2017/11 - 2019/09) - Automation of SQL-Reportings - Implementation of self-service BI-tool Qlik Sense - Design and implementation of ETL processes for the existing data warehouse and the Qlik self-service architecture - Technical administration of the Qlik Sense Server - Data analysis using statistical as well as machine learning methods with R and python (cluster analysis, natural language programming)

- Adhoc reporting in SQL (Oracle) and Excel for various departments - Analysis of CRM newsletter campaigns - Support in PoC for a self service BI-Tool

Masterthesis: Instance Segmentation of Holograms on ID Cards - Python - Tensorflow/Keras - Face Detection (MTCNN) - Mask R-CNN / U-Net as main model