Created an innovative Deep Learning-driven process monitoring system employing LSTM-Autoencoder technology. The system excels in real-time anomaly detection within multivariate sensor data, significantly enhancing operational availability and preventing potential damage to valuable workpieces.

Innovated two cutting-edge loss weighting methods for Multi-task Learning (MTL) that surpassed existing state-of-the-art approaches. Successfully integrated these novel methods into the established codebases of Bosch and Cariad (corporate partner) for automated driving applications. Conducted comprehensive evaluations of loss balancing techniques from the literature across various tasks, including semantic segmentation, object detection, depth, and surface normal estimation, using scene understanding datasets such as Cityscapes and NYUv2. My work resulted in the registration of the two groundbreaking loss weighting methods as patents, and it also paved the way for the submission of two papers to the prestigious scientific conference CVPR (2023 and 2024).

Developed processes for loading the staging area and the raw data vault by employing AWS services such as S3, Lambda, and Batch. Led the internal BI development team as a Scrum Master.

Conducted image-to-image translation between the domains of regular images and artworks with Deep Generative Adversarial Networks. Enhanced CycleGAN by introducing a two-objective discriminator as regularization, incorporating an adversarial self-defense for better cycle-consistency, and applying differentiable augmentation on the target domain with less data. Employed agile intercultural project management techniques to manage the project successfully.