You would join a team and a project based under the AI-FES (AI in Feature Effort Support) registered with the MN RAN AI Project Steering Team. The focus of this internship would be recommending code files based on historical features, similar to a new feature intended for development, by the use of the neural code search approach.

Position: AI FES - Working Student
Duration: 6 months with possibility to extend
Weekly Time Allocation: 10-20 hours
Location: Ulm, Germany
Start Date: 15.June 2025
Education Recommendation: Preferably Master's student or above
 

You have:

Experience in working with LLMs and embedding models, preferably fine tuning of said models Ability to create Front-end for AI based applications Familiarity with the AI lifecycle i.e. from data collection through to model retraining based on external/pilot feedbacks Working experience with integrating RAG on top of given AI application

 

It would be nice if you also had:

Knowledge on evaluation criteria and metrics for AI models against given use cases

 

As part of our team:

you will be looking into building an embedding model that will fuse code to the related requirements using Contrastive Training Scheme or similar approach. You will get an opportunity to integrate this model into a Proof of Concept (PoC) where we want to find what type of requirements effect which parts of the code. As part of the PoC, we will study how to proceed with this concept, and what other mechanisms exist to reach the intended goal.