Director of Machine Learning / Principal Scientist
since 2022
co:here
I'm leading the team that develops & trains Large Language Model (up to 100B+ parameters) for semantic search and few-shot text classification. Our goal is to develop foundation models that can create rich content representations for many use-cases & tasks.
Research Scientist
2021 - 2022
HuggingFace
I build the Neural Search research team at HuggingFace from scratch. We developed methods that allow to use Neural Search without the need of labeled training data. I'm the creator and core maintainer of sentence-transformers.
Post-Doctoral Researcher
2018 - 2021
Technical University of Darmstadt, Germany
As post-doctoral researcher I am leading different research projects and work jointly with my Ph.D. students to achieve new research contributions. The projects focus on using deep learning in the area of Natural Language Processing (NLP). I worked as a lecturer for the class Deep Learning for Natural Language Processing. See the research for more details.
Doctoral Researcher
2014 - 2018
Technical University of Darmstadt, Germany
I was working as a researcher in the UKP Lab at the Technical University of Darmstadt. I worked on various research projects connected to deep learning and Natural Language Processing (NLP). I supervised full-time research assistants and master students. I worked as teaching assistent. See the research for more details.
Fellow Intern
2013
MyKinsey & Company
Experiences in the field of a lean IT and IT consolidation with a focus on Enterprise Architecture Management.
Founder
I have founded several companies, alone as well as with co-founders. I specialized in the creation of highly automated online businesses that require minimal labor to run.
2018
WebsiteUptime.de is a software-as-a-service that monitors the availability of websites and servers. In case of issues, a notification is send to the owner. A couple of thousands of active customers.
WebhosterWissen.de 2016
WebhosterWissen.de continuously measures the quality of various web hosting services, virtual private servers and cloud servers. Reports are generated semi-automatically comparing the quality of the different providers. About 50,000 visits / month.
2014
Stipnetz is the largest social network of scholarship students in Germany, Austria and Switzerland. The underlying software was custom build with PHP (Laravel), MySQL and React to meet the special needs of the users. As a CTO I coordinated a team of five software developers and web designer.
digicount (co-founder & CTO)
2005
Digicount was a subscription-based web analytics service that tracked and reported website traffic. With the live reports website owners could gain insights on their website visitors and optimize their website accordingly.
PHP-Einfach.de 2003
PHP-Einfach.de is an interactive online-course to learn software development with PHP & MySQL. It is still active today and has more than 200,000 visits per month.
Freelancing - AI Constulant & Machine Learning Engineer
Since 2015 I consult companies on their AI strategy, offer trainings for the most recent AI technologies, and provide software development skills for the creation of state-of-the-art machine learning solutions. My clients are in the area of investment banking, publishing houses, and Internet-of-Things. See my portfolio for more details.
Freelancing - Web Development
Since 2003 I work as a full-stack web developer. For various clients, I developed web applications mainly using PHP (Laravel), MySQL, and JavaScript (React). Project scopes rang from small applications (few weeks of development) up to large scale applications with over 12 months of development and various stakeholders. See my portfolio for more details.
Education
Ph.D. - Natural Language Processing
2014 - 2018
Technical University of Darmstadt, Germany
I finished in 2018 my Ph.D. summa cume laude in the area of deep learning and natural language processing. I pursued my Ph.D. under the supervision of Prof. Iryna Gurevych at the UKP Lab at the Technical University of Darmstadt.
Master of Science - IT-Security
2011 - 2014
Technical University of Darmstadt, Germany
In 2014 I graduated from the TU Darmstadt with a Master of Science in IT-Security. During the master program I especially focus on algebraic cryptography.
Semester Abroad
2010
University of California, San Diego, USA
Semester Abroad
2009
University of California, Berkeley, USA
Bachelor of Science - Computer Science
2007 - 2011
University of Oldenburg, Germany
Bachelor of Science - Mathematics
2007 - 2011
University of Oldenburg, Germany
Top Repositories
15601
2268
2031
1657
1183
Selected Publications
- GPL: Generative Pseudo Labeling for Unsupervised Domain Adaptation of Dense Retrieval. 2021. [pdf] [git]
- BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models. 2021. [pdf] [git]
- Augmented SBERT: Data Augmentation Method for Improving Bi-Encoders for Pairwise Sentence Scoring Tasks. 2020. [pdf] [git]
- Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation. 2020. [pdf] [git]
- Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. 2019. [pdf] [git]