RAG systems often perform well in principle, but real-world use cases require higher robustness and quality control. In Part III of this workshop series, participants learn how to systematically improve RAG systems using Reinforcement Learning from Human Feedback (RLHF).
The workshop covers feedback dataset creation, evaluation strategies, and the optimisation of retrieval versus generation components. Through practical exercises, participants evaluate real RAG outputs, run optimisation loops, and observe measurable improvements in system performance.
Part III connects technical RAG design (Parts I & II) with the development of production-ready assistant systems (Part IV).
Event Details
-
Title: Efficient Information Retrieval – Part III
-
Format: In-person workshop
-
Topic: Reinforcement Learning from Human Feedback for RAG
-
Date: Tuesday, May 19, 2026
-
Time: 10:00 AM – 1:00 PM (CEST)
-
Location: Hasso-Plattner-Institut
Prof.-Dr.-Helmert-Straße 2–3, Building K, Room K 1.04, 14482 Potsdam -
Organizer: AI Service Center Berlin-Brandenburg @ HPI
-
Duration: 3 hours
-
Admission: Free, registration required
-
Requirements: Participation in Parts I & II or equivalent experience with RAG systems
-
Note: Photo, audio and video recordings will take place during the event
Hass-Plattner-Institut
Prof.-Dr.-Helmert-Straße 2–3, Building K, Room K 1.04, 14482 Potsdam