Starbucks A/B Testing project

Analysis of an A/B test of a customer promotion. Assessing the significance of the impact on incremental response rate (IIR) as well as net incremental revenue (NIR). Improving the targeting of customers to further increase IIR and NIR by developing models based on customer data.

Disaster Response Pipeline

Implementation of a response pipeline for twitter posts. Building a web app that uses a machine learning model to predict disaster categories from input tweets for fast response. Development of an ETL pipeline, a machine learning pipeline and deployment of model to a personal website using docker.

IBM Watson Studio Recommendation Engine

Building a recommendation engine for articles from the IBM Watson Studio platform. Implementation of collaborative filtering and content-based recommendation methods. Developing method for recommending articles to new users using hybrid approach combining Latent Dirichlet Allocation (LDA) and GPT-3.

Tutorial: Deep Learning with PyTorch

Developing and teaching of an exercise for students at the university of Bonn. Implementing and demonstrating how to build deep neural networks in PyTorch and use them for classification, regression and computer vision tasks.

Tutorial: Normalizing Flows with TensorFlow

Developing and teaching an exercise at the 2nd Terascale School of Machine Learning in 2021, taking place at DESY in Hamburg. Providing a theoretical introduction into normalizing flows and showing how to implement and apply them to a particle physics problem using the TensorFlow-probability package

Tutorial: Unsupervised Anomaly Detection in Particle Physics

Developing and teaching an exercise at the 3rd Terascale School of Machine Learning in 2022, taking place at DESY in Hamburg. Writing a theoretical introduction about resonant anomaly detection problems in particle physics. Describing and implementing two methods for unsupervised anomaly detection: