Deep learning might currently be one of the few technologies that has a great transformative potential for us all. To advance deep learning, I think one of the most important aspects is to understand the technology very well. Such an understanding should be accessible to as many people as possible. Thus, I am excited whenever I have the chance to work on something that fosters this understanding or makes it available to more users.

Work

Founding Member of Technical Staff (May 2024 - Present)

AI + Biology

Postdoctoral Researcher, Carnegie Mellon University (Oct 2023 - May 2024)

Research and development of tools in the areas of AI evaluation and prompt engineering.

Research Scientist, Sigma Computing (Oct 2022 - Jun 2023)

Bringing data and analysts closer together with the help of visualization and AI.

Research Intern, Apple Machine Intelligence (Mar 2021 - Sep 2021)

Designed and developed a framework for component-based ML interfaces which can be composed in different environments such as computational notebooks and web dashboards.

Research Intern, Google TensorBoard (Jun 2020 - Sep 2020)

Designed and implemented a visualization approach for a novel bias detection algorithm. This visualization is designed to support large label spaces and multilabel classification problems.

Research Intern, Google PAIR (May 2019 - Aug 2019)

Invented and experimented with a technique similar to Feature Visualization, but for language models.

Talks

Unlocking Model Performance Insights with Zeno

Multimodal Weekly, (Fri Jan 26 2024)

Where did my Lines go? Visualizing Missing Data in Parallel Coordinates

EuroVis 2022, Rome, Italy, (Wed Jun 15 2022)

Symphony: Composing Interactive Interfaces for Machine Learning

ACM CHI 2022, New Orleans, LA, USA, (Wed May 04 2022)

Net2Vis - A Visual Grammar for Automatically Generating Publication-Tailored CNN Architecture Visualizations

IEEE VIS 2021, New Orleans, LA, USA (remote), (Thu Oct 28 2021)

Visualization for AI in Critical Domains

Bio+Med+Vis Spring School, Brno, Czech Republic (remote), (Mon Oct 26 2020)

What Does BERT Dream Of?

VISxAI 2020, Salt Lake City, UT, USA (remote), (Mon Oct 26 2020)

Classifier-Guided Visual Correction of Noisy Labels for Image Classification Tasks

EuroVis 2020, Norrköpping, Sweden (remote), (Wed May 27 2020)

Automatic Fibril-Crossover Detection in EM-Images using Deep Convolutional Networks

MedVis Workshop 2018, Ulm, Germany, (Thu Apr 12 2018)

Teaching

Lecture: Deep Learning for Graphics and Visualization (2019 - 2022)

Created and regularly held one chapter on visualization for AI at Ulm University. Talk about different explainability techniques and visualization concepts that help investigate and communicate about AI systems.

Projects: Visualization and Explainability for AI (2017 - 2022)

Supervise undergrad and grad students in regular projects. We go through the process of ideation, implementation, and writing about these projects.

Seminars: Visualization and Explainability for AI (2017 - 2022)

Held regular seminars for both undergrad and grad students.

Reviewing and Service

VGTC - Industrial Relations Chair (2023 - Present)

VISxAI - Program Committee (2022 - Present)

TVCG (2020 - 2023)

VIS (2021, 2023)

CHI (2021, 2023)

UIST (2022)

CG&A (2022)

EuroVIS (2022)

IMWUT (2022)

VCBM (2022)

MICCAI (2022, 2023)