Daniel Handloser

• Implemented a component to extract HTTP status codes via Selenium, which identified that 14% of GUI interactions lead to errors. Added 30+ posts to internal wiki on machine learning. • Designed and implemented a Java tool that exports training data from existing test scripts and user traces. Created the company’s deep learning training data based on that.

• Improved graph visualization in a forensics tool, reducing 62% SLOC in the visualization package in Java • Developed programs for data visualization and created a chat bot for customer interaction

CS Master’s thesis: Trained and evaluated Neural Response Generation models for less bland outputs. Based on pre-trained Transformer and a hierarchical recurrent model in Python with PyTorch and Tensorflow.

• Automatic Speech Recognition, Cloud Computing

• Deep Learning, Machine Learning I + II, Deep Learning for Computer Vision, Human-Machine-Interaction • Minor in Entrepreneurship

• Bachelor’s thesis: Predicting the stock market using neural networks. Created models in Keras and Python that predict NASDAQ and DAX up and down changes with 65% accuracy. • Minor in Finance