Welcome @stardust 
Make sure to spend time looking at all the data and tools to come up with new investigations. Looking at previous studies is quite useful too.
Previous work: NASA OSDR-GeneLab Enabled Publications, Pre-Prints, Student Theses - Google Scholar
Data dashboard (at a glance): NASA OSDR Biological Data and Visualization Portal
Search portal: NASA OSDR: Open Science for Life in Space
BioData API: https://visualization.osdr.nasa.gov/biodata/api/
Our code on GitHub: OpenScienceDataRepo · GitHub
Training courses (GL4U and AI/ML for space biology): https://www.nasa.gov/reference/gl4u-training-resources/
These and all quick links to data/tools/APIs, etc are here: About - Open Science Data Repository Analysis Working Groups
RAG and NLP are a good fit here and mostly untouched. OSDR has hundreds of studies with rich structured metadata, +200 publications built on that data, and nobody has fully worked out how to get an LLM to reason across both. The BioData API gives you the metadata programmatically, so a good first exercise is to pull it, see how it is structured, and decide what you would build on it.
The members who get the most out of the AWGs show up with something concrete: a dataset they have already looked at, a question, or a small prototype. Spend a couple of weeks in the data, then bring what you find to the AI/ML monthly meeting on the calendar (log in first). If it grows into a project idea, propose it in the master table thread so others can join you.
Cheers,
-Ryan