Cognitive Scientist, Senior DevOps Engineer
I am a cognitive scientist turned programmer seeking a career change. At the University of Bremen, I gained more than three years of academic experience in natural language understanding for household robots. I then joined the infrastructure team at Fraunhofer MEVIS, where I served as deputy team lead and built and maintained the institute’s deep learning infrastructure as a Senior DevOps Engineer. Now I am looking forward to pursuing new challenges.
Fraunhofer-Gesellschaft
Fraunhofer Institute for Digital Medicine MEVIS
– Present
At MEVIS I am responsible for conceptualizing, configuring, and maintaining the institute's edge cloud computing platform based on the HashiCorp stack. As deputy team lead, I am involved in strategic decision-making, job interviews, and task prioritization. My duties also include internal consulting on CI/CD. Additionally, I work on the MeVisLab build infrastructure and third-party integration, including CVE monitoring and SBOM generation.
Fraunhofer-Gesellschaft
Fraunhofer Institute for Digital Medicine MEVIS
–
Edge computing platform configuration, automation of manual tasks, bridging the gap between classical IT, DevOps, and users. Deep learning and service infrastructure conceptualization, configuration, and maintenance: Hashicorp Stack, GitLab CI/CD.
University of Bremen
Digital Media Lab
–
In the DFG-funded collaborative research center "EASE", we aimed to build models for everyday activities and deploy them on household robots. I worked on natural language understanding and built a pipeline to transform natural language instructions into action plans for robots. Additionally, my lab colleagues and I worked on examining design decisions of questionnaires in VR and detecting a break in presence when taking off VR headsets.
Osnabrück University
Institute of Cognitive Science
–
As a lecturer, I taught "Basic Programming in Python", a course tailored to Master students with a non-programming background. My teaching performance was rated in the top 10 % of all lectures that semester.
Osnabrück University
Institute of Cognitive Science
–
At the Institute of Cognitive Science, I was a tutor for Machine Learning and Computer Vision for three semesters. I reviewed homework assignments, tutored students and rewrote and designed exercises in Jupyter notebooks.
Level Up Analytics GmbH
Intuit Data Engineering and Analytics
–
At Level Up Analytics, I worked on open sourcing Wasabi, Intuit's internal A/B testing SaaS. Additionally, I developed features for the Wasabi's backend and resolved various issues with its internal ETL pipeline passing data from Hadoop to Vertica.
Intuit Inc.
Intuit Data Engineering and Analytics
–
I added custom tailored filter and bookmarking features for Wasabi's administration UI. Also, I worked on a docker setup for Windows and delivered a talk on machine learning.Also, I worked on docker images of Wasabi for Windows.
Intuit Inc.
Intuit Data Engineering and Analytics
–
During my internship at IDEA, I worked on the Wasabi team. I added an audit log to Wasabi and delivered an introductory talk on Machine Learning.
Master of Science · GPA: 1.1 (with distinction)
Osnabrück, Germany–
Thesis: Gaze Tracking Using Common Webcams
Master of Science (discontinued)
Osnabrück, Germany–
Bachelor of Science · GPA: 1.9
Osnabrück, Germany–
Thesis: Probabilistic Robot Localization in Continuous 3D Maps
Erasmus Exchange · GPA: 1.25 (converted)
Ankara, Turkey–
High School · GPA: 1.8
Sundern, Germany–
Inside (school magazine)
–
Layout and editing of articles, organizing team meetings.
2022
Going from natural language directions to fully specified executable plans for household robots involves a challenging variety of reasoning steps. In this paper, a processing pipeline to tackle these steps for natural language directions is proposed and implemented. It uses the ontological Socio-physical Model of Activities (SOMA) as a common interface between its components. The pipeline includes a natural language parser and a module for natural language grounding. Several reasoning steps formulate simulation plans, in which robot actions are guided by data gathered using human computation. As a last step, the pipeline simulates the given natural language direction inside a virtual environment. The major advantage of employing an overarching ontological framework is that its asserted facts can be stored alongside the semantics of directions, contextual knowledge, and annotated activity models in one central knowledge base. This allows for a unified and efficient knowledge retrieval across all pipeline components, providing flexibility and reasoning capabilities as symbolic knowledge is combined with annotated sub-symbolic models.
2021
In this paper, we present foundations of the Socio-physical Model of Activities (SOMA). SOMA represents both the physical as well as the social context of everyday activities. Such tasks seem to be trivial for humans, however, they pose severe problems for artificial agents. For starters, a natural language command requesting something will leave many pieces of information necessary for performing the task unspecified. Humans can solve such problems fast as we reduce the search space by recourse to prior knowledge such as a connected collection of plans that describe how certain goals can be achieved at various levels of abstraction. Rather than enumerating fine-grained physical contexts SOMA sets out to include socially constructed knowledge about the functions of actions to achieve a variety of goals or the roles objects can play in a given situation. As the human cognition system is capable of generalizing experiences into abstract knowledge pieces applicable to novel situations, we argue that both physical and social context need be modeled to tackle these challenges in a general manner. The central contribution of this work, therefore, lies in a comprehensive model connecting physical and social entities, that enables flexibility of executions by the robotic agents via symbolic reasoning with the model. This is, by and large, facilitated by the link between the physical and social context in SOMA where relationships are established between occurrences and generalizations of them, which has been demonstrated in several use cases in the domain of everyday activites that validate SOMA.
2020
Questionnaires are among the most common research tools in virtual reality (VR) user studies. Transitioning from virtuality to reality for giving self-reports on VR experiences can lead to systematic biases. VR allows to embed questionnaires into the virtual environment which may ease participation and avoid biases. To provide a cohesive picture of methods and design choices for questionnaires in VR (inVRQ), we discuss 15 inVRQ studies from the literature and present a survey with 67 VR experts from academia and industry. Based on the outcomes, we conducted two user studies in which we tested different presentation and interaction methods of inVRQs and evaluated the usability and practicality of our design. We observed comparable completion times between inVRQs and questionnaires outside VR (nonVRQs) with higher enjoyment but lower usability for inVRQs. These findings advocate the application of inVRQs and provide an overview of methods and considerations that lay the groundwork for inVRQ design.
2020
Questionnaires are among the most common research tools in virtual reality (VR) evaluations and user studies. However, transitioning from virtual worlds to the physical world to respond to VR experience questionnaires can potentially lead to systematic biases. Administering questionnaires in VR (inVRQs) is becoming more common in contemporary research. This is based on the intuitive notion that inVRQs may ease participation, reduce the Break in Presence (BIP) and avoid biases. In this paper, we perform a systematic investigation into the effects of interrupting the VR experience through questionnaires using physiological data as a continuous and objective measure of presence. In a user study (n=50), we evaluated question-asking procedures using a VR shooter with two different levels of immersion. The users rated their player experience with a questionnaire either inside or outside of VR. Our results indicate a reduced BIP for the employed inVRQ without affecting the self-reported player experience.
2019
In this paper, we introduce the framework of MEANinGS for the semi-autonomous accumulation of world knowledge for robots. Where manual aggregation is inefficient and prone to incompleteness and autonomous approaches suffer from underspecified information, we deploy the human computation game Kitchen Clash and give evidence of its efficiency, completeness and motivation potential.