Open Postions for Student Research Assistants
In collaboration with the High Performance Computing research group of Prof. Plessl, we offer positions for student research assistants (Studentische Hilfskräfte SHK/WHB). Typical activities include contributions to research and development projects and support for courses.
Current offers can be found below.
Application process
To ensure prompt processing of your application and feedback, we ask you to submit a meaningful application comprising the following information:
- Cover letter stating your interest and previous knowledge for the desired topic
- Curriculum vitae with an overview of your university education and major fields of study.
- List of university courses taken so far and certificates
- Experience relevant to the position, such as professional activities or student projects, personal projects such as contributions to open source projects, etc.
Please send your application directly to the contact person for the topic.
Currently open positions
Performance testing of LLM inference frameworks
Description
In a joint project with ZIM and Hochschuldidaktik, we are building a prototypical infrastructure for the use of Large Language Models in teaching. For this, we are looking for student assistants to support us in evaluating and implementing the server backend for LLM inference.
The task includes planning and executing the evaluation of different LLM inference engines, such as ollama/llama.cpp, vllm, Nvidia Triton, LMDeploy, MLC-LLM, TensorRT-LLM, or tgi. The result should be a consistent overview of the features and performance on the GPU types available on PC2 (NVidia A100, A40, H100) and AMD (MI 250).
The following benchmarks are a good starting point for the work:
- https://github.com/ninehills/llm-inference-benchmark
- https://www.bentoml.com/blog/benchmarking-llm-inference-backends
- https://blog.premai.io/prem-benchmarks/
Desired Skills
For this task, knowledge in the following areas is desirable or required:
- Experience with language models from a programmer's point of view (API use)
- Overview of the requirements of LLM inference
- Good knowledge of Python scripting
- Good knowledge of using Linux systems
- Ability to document evaluation results in a more comprehensible way
The following skills are an advantage but not essential:
- Use of software containers: Docker, Singularity
- Experience with using the HPC systems of PC2
Contact
Your contact person: Robert Schade
Keyword: "LLM performance testing job offer"
Apply via e-mail
Software development for self-service and management infrastructure
Description
The dynamic environment of an HPC center requires custom software solutions. To meet all the requirements by employees as well as users, the PC2 provides both a user portal, containing self-services for users, and PERSEUS, a management software for compute time proposals and compute projects. To increase our capacities and enhance our provided services, we are looking for student assistants to support our software development team.
The task includes working on different software projects by implementing required functionalities with Python and / or Typescript and React, implementing automated tests and writing documentation. This also requires working with third-party software and their corresponding interfaces. Additionally, it includes further developing already existing CI/CD pipelines and optimizing the deployment strategies.
Desired Skills
For this task, knowledge in the following areas is desirable or required:
- Experience with Python and / or TypeScript in combination with React
- Knowledge of CI/CD pipelines
- Experience with REST APIs
The following skills are an advantage but not essential:
- Experience with Docker containers
- Experience with MongoDB
- Experience with FastAPI or a similar Python framework
- Experience with automatic software testing
Contact
Your contact person: Lukas Ostermann
Keyword: "software development job offer"
Apply via e-mail
Technical Editing, Debugging and Infrastructure Assistant
Description
We work on topics generally related to Go, an ancient, Asian boardgame that is still played by a huge community today.
Go has been a major driver of advancements in game theory, decision making under uncertainty, and AI research.
Computer Go recently had a breakthrough with the advent of Deepmind's AlphaGo, which for the first time in history could beat a human professional player.
For this purpose, it employed Deep Convolutional Neural Networks in combination with Monte Carlo Tree Search, a type of informed search that is stupidly simple but works surprisingly well.
There is an abundance of aspects in Go (or similar games) that are actively researched.
If you enjoy this work and want to dive in deeper, we can also find an interesting topic for your BSc or MSc thesis.
Desired Skills
Tasks may include one or more of the following, depending on what you like best.
We can work out your strengths and focus on them!
No matter which aspect you opt for, it will directly or indirectly allow you to gain or improve skills that are helpful in your scientific career.
Linguistic improvements of English technical texts
Editing LaTeX documents, e.g. layout, bibliography, formatting, troubleshooting
Enhancement of graphics created with Matplotlib or Seaborn
Conversion, validation, or basic analysis of measurement data
Programming in Python using e.g. Pandas, Scipy, Pytorch
Programming in C++ using e.g. OpenMPI
Modifications to KataGo, the leading open source Go program
Writing helper programs or scripts in a linux environment using bash, Python or C#
Documentation of and adjustments to infrastructure and workflows
Requirements:
Conscientious working style
Fluent German, spoken and written
Good written English skills
Knowledge of programming (any language) and/or LaTeX
Contact
Your contact person: Sebastian Heuchler
Keyword: Technical Editing, Debugging and Infrastructure Assistant
Apply via e-mail
Paderborn Center for Parallel Computing (PC2)
Mersinweg 5
33100 Paderborn