Jonathan Decker

E-mailjonathan.decker@uni-goettingen.de

Biography

Jonathan is a scientific employee of the Georg-August-University of Göttingen and a PhD student of Julian Kunkel.
He takes the role of a system architect and is focused on designing systems that enable new and novel ways of utilizing Cloud and HPC resources, while also being efficient, secure and scalable. Most notably, he strives to combine HPC with Kubernetes.
ORCID: 0000-0002-7384-7304

Research Interests

  • Kubernetes
  • Serverless

Teaching

Autumn Term 2024

Summer Term 2024

Autumn Term 2023

Summer Term 2023

Autumn Term 2022

Summer Term 2022

Open Thesis Topics

Fixing Shortcomings of Kubernetes Severless TechnologiesApply

Serverless Computing or Function-as-a-Service (FaaS) has emerged as a new paradigm for computing over the last few years. There exists a number of open source FaaS platforms based on Kubernetes as the container orchestration platform maps well to the components required for FaaS. However, most approaches to FaaS are still relatively naive and leave many performance improvements on the table. This work focuses on said limitations and aims to solve at least one of them and implement a proof of concept. Finally, the performance improvements should be benchmarked in a virtualized environment and on the HPC system.

Jonathan Decker HPCContainer

Confidential GPU InferenceApply

For customer facing systems that handle sensitve data such as patient information, it is required to comply with strict data protection laws. In order to comply with these laws even during a security breach, confidential computing should be used, however, modern use-cases require the usage of scable multi-user systems with GPU acceleration for ML inference workloads. This thesis encapsulates setting up confidential computing on top of a Kubernetes cluster using Kata Containers, Confidential Containers and Nvidia Confidential GPU Computing as well as measuring the performance costs of using a confidential compute stack.

Jonathan Decker MLContainer

Advancing Source References in RAG LLM SystemsApply

Retrieval-Augmented-Generation (RAG) is a method for providing an LLM with additional data via documents that are automatically added into queries. This method has been implemented in a number of ways including multiple open source projects such as PrivateGPT or H2OGPT. These projects commonly have a feature where the answer to a given query states what documents and what pages in these documents were used to complete the query, however, these can often be inaccurate as the system is forced to make a selection even in cases where no document has been used. For this thesis topic, a student would set up such a RAG LLM system and augment its retrieval system to provide more accurate references to documents or none if no documents from the given set were used.

Jonathan Decker AIGPT

Theses

  • The Potential of Serverless Kubernetes-Based FaaS Platforms for Scientific Computing Workloads, Jonathan Decker (Master's Thesis), Advisors: Dr. Julian Kunkel, 2022-01-31, BibTeX URL

Publications

2024

2023

  • DECICE: Device-Edge-Cloud Intelligent Collaboration Framework (Julian Kunkel, Christian Boehme, Jonathan Decker, Fabrizio Magugliani, Dirk Pleiter, Bastian Koller, Karthee Sivalingam, Sabri Pllana, Alexander Nikolov, Mujdat Soyturk, Christian Racca, Andrea Bartolini), ACM, Computing Frontiers, ISBN: 979-8-4007-0140-5/23/05, 2023-05-09 BibTeX DOI PDF

2022

All publications as BibTex