Ricardo Chavez

(Chair of Human-Machine Communication) I investigate various methods for creating Audio Adversarial Examples (AAE's) in an end-to-end Automatic Speech Recognition system called ESPnet. The framework is based on RNN (BLSTM) networks and is implemented using PyTorch. I then use the AAE's information to increase the model robustness by 24-33% absolute on the AAE's augmented set employing Adversarial Training, which actively infuses AAE's into the training process.

General manager of Die Bierstube, part of the Association of Students in Olympiazentrum. The job included leading an international team of students, recruiting new employees, taking care of supplies and food orders, and implementing new projects to improve student life.

Assistance in office and electronics laboratory, implementation of firmware updates and customer-specific parameterization of various radio modules and devices (wireless technologies, such as BLE, Beacon, IoT), and development of a project-specific module of the Xentral ERP system for printing special material labels

The program "adveisor" is designed for 1st and 2nd-semester students, where students can translate their electrical engineering knowledge into a project and improve their soft skills. I took part in seminars and workshops on team leadership and development, communication, project, and time management. I was able to implement the knowledge acquired to guide a group of students and together obtain the 1st place in the program.

Chair of Data Processing: Designing and implementing a digital lending system for traineeship materials, including the creation of a Graphical User Interface with Qt Quick

Chair of Human-Machine Communication: Tutor for the course "Signal Processing" and "System and Analog Circuits Laboratory" and Improvement of the MATLAB-homework for the course "Signal Processing"

Automation and Robotics. Thesis Topic: Improving Robustness of Sequence-to-Sequence Automatic Speech Recognition by Means of Adversarial Training

Machine learning and Artificial intelligence Thesis: Human-robot team shared control for cooperative manipulation tasks