MunichNLP x Tech’n’Drinks @ ALL IN GROUP - Building RAG systems

MunichNLP x Tech’n’Drinks @ ALL IN GROUP
Building reliable, explainable & efficient RAG systems

Hello Munich NLP Enthusiasts!
We are excited to announce our next joint meet up with Tech’n’Drinks, taking place on Thursday, May 7, 2026 at the MYPOSTER office of **ALL IN GROUP **at Stiglmaierplatz in Munich!
Join us for an evening of talks, discussion, and networking at the intersection of industry, research, and real-world AI systems. This edition features a joint talk by Andreas Stephan and Matthias Aßenmacher on data science projects across legal industry and science, followed by a talk by Andrei Beliankou on building reliable dialog graphs for LLM-based systems.

🕵️‍♀️ How to find: The event takes place at the MYPOSTER office of ALL IN GROUP, Nymphenburger Straße 12, 80335 Munich, right at Stiglmaierplatz
📅 Agenda
18:00 | Doors open + 🍴 food & drinks
18:45 – 19:00 | Intro (Sponsor, Tech’n’Drinks & MunichNLP)
19:00 – 19:40 | Talk 1:“The Data Science Project – Perspectives from the Legal Industry and Science” Andreas Stephan & Matthias Aßenmacher
19:45 – 19:50 | Short break
19:50 – 20:30 | Talk 2: TBA Andrei Beliankou
20:30 – 21:30 | Networking & drinks

🔎 First Talk: “The Data Science Project – Perspectives from the Legal Industry and Science”
Speakers: Andreas Stephan and Matthias Aßenmacher
Time: 19:00 – 19:40
Abstract
In the abstract, the core data science project often seems remarkably similar whether you're working on a product or on an academic research project. In this talk we compare multiple projects from the viewpoints of industry and academia. Specifically, we compare a retrieval augmented generation (RAG) project in the legal industry, an applied research project using RAG to analyze market research papers, and a pure research study investigating the LLM-as-a-Judge paradigm. In the process we will discuss similarities and differences in problem definitions, methodologies, and evaluations.

About the Speakers
Dr. Andreas Stephan is Tech Lead at the legal tech startup LDA - Legal Data Analytics GmbH. He holds a PhD in NLP from the University of Vienna and brings industry experience from his previous work as an NLP data scientist in the finance and insurance sectors

Dr. Matthias Aßenmacher postdoctoral researcher at the Chair of Statistical Learning and Data Science at LMU and the NFDI Consortium for Business, Economic and Related Data (BERD@NFDI). He focuses on teaching NLP and works on a diverse set of NLP Applications, including Uncertainty/Subjectivity, NLP for historical/low-resource languages, Bias in NLP, and NLG/Decoding from LLMs.

🔎 Second Talk: “TBA”
Speaker: Andrei Beliankou
Time: 19:50 – 20:30
Abstract
Moving a RAG system from prototype to production quickly reveals challenges that go far beyond prompt design and retrieval quality. In this talk, Andrei shares practical lessons learned from operating RAG systems in real environments, with a focus on MLOps, system availability, redundancy, observability, and reliability. He will discuss common failure modes, trade-offs in architecture and deployment, and what it takes to build systems that remain stable, explainable, and cost-efficient under real-world constraints.

About the Speaker
Andrei Beliankou is Technical Lead Data & AI at E.ON Digital Technology in the Energy Retail Team. As part of the E.ON GenAI Core team, he has been developing generative AI solutions for different business units across the organization. His work focuses on areas such as multilingual search, LLMs, agentic RAG, and knowledge graphs. He describes himself, unofficially, as a software engineer with a strong data affinity, writing both code and texts.

🍴 Food & Networking
Before the talks food & drinks will be served. After the talks, we will continue with drinks, and networking. As usual, this is a chance to connect with others from the local NLP, data science, and GenAI community across academia and industry.