Qualification portal for PhD students

* Building Your Own AI Research Agent: Translating Hermes, HPC and SAIA into Practical Tools for Everyday Research (GGG) online

AI agents can take over recurring, time-consuming research tasks while giving researchers greater control over the models, data, and infrastructure they use. This session translates a technical setup built with Hermes, GWDG HPC resources, and the SAIA API into an accessible blueprint that researchers can understand, adapt, and build upon.

Target group:
PhD students of GGG with varying levels of AI experience, other PhD students if free places are available

Schedule:
09 November 2026
12:00 to 13:30 h

Venue: online (You will receive login data via e-mail in advance.)
Available seats: 12
Language: English

This is the fifth of six AI Sessions of GGG. For an overall description of the series see https://www.uni-goettingen.de/en/707971.html

Trainer: The interactive online sessions are facilitated by Dr. Christian Steinau. Christian Steinau enjoys formats in which dialogue leads to concrete action. His academic home is LMU Munich, where he earned his doctorate on the topic of judgment: a skill that is becoming increasingly important in today’s work with AI. After all, the better AI systems become at producing texts, analyses, and ideas, the more important it becomes to consider how we evaluate, contextualize, and responsibly utilize their outputs. As the director of a Transfer Lab, he worked at the intersection of research, society, and application. In 2020, he founded the Cultural Policy Lab, which he has been leading since 2025 as a research firm specializing in the cultural and creative industries. His initiatives combine a curiosity about new technologies with scientific rigor, peer learning, and a clear focus on integrity and practical value.

Internal experts: In addition to the trainer, experts from the GWDG will be available for questions and exchange on internal AI tools and services.

Participants are warmly invited to contribute their own questions, experiences, tools, workflows and critical perspectives. The aim is to learn from one another, build networks across disciplines, and develop practical, reflective and future-oriented ways of working with AI in doctoral research.
In this session, the trainer will walk you through his own build process step by step:

  • The Setup: How the agent was conceptualized and assembled, including how HPC access was set up, how the SAIA API was connected, how Hermes was configured, and how the different components were brought together into a working research environment.
  • Demonstration: Seeing the agent in action (for example, searching for relevant literature, tracking funding opportunities, and monitoring conferences and publications).
  • Translation plus Q&A: Breaking down the technical components into understandable concepts and discussing how they could be adapted to different research projects.

You will also look at how different models can be integrated, and what this can mean for greater control over research data and infrastructure.

You do not need to be an HPC expert to benefit from this session. No prior experience is required, although parts of the session will get a little technical. The aim is precisely to translate this technical infrastructure into concepts that researchers can understand and use.

By the end of the session, participants should understand the basic architecture well enough to start setting up an agent for themselves or their own research project. Documentation and step-by-step guidance will be provided. Participants can use this to explore applying for HPC resources for their own research project or, where available, use test access provided through the GGG AI Session.

The goal is not to turn everyone into an HPC specialist, but to provide enough understanding and practical guidance to start experimenting with AI agents as an emerging part of research infrastructure.

Credits: 0 (You will receive a proof of attendance.)
Requirements:

  • full participation in the entire session (camera on)
  • active participation in discussions and exchange


Registration:
Please, register online via the registration form for GGG courses.
In the registration form, under course-specific information, please indicate...

  • your expectations for the session
  • specific questions / challenges you would like to discuss
  • if you would like to present something

Please also note our regulations on bindingness: e.g. four weeks before a course starts, all course registrations are regarded as binding.

Contact for further information:
Dr. Nelly C. Schubert
Phone: +551 39-28219
E-mail: ggg.kursanmeldung@uni-goettingen.de

This session is organized by the Göttingen Graduate School of Social Sciences (GGG).

GGG Logo 2016