15th Munich NLP Meetup feat. LLMs with deepset & dstack

Hey everyone,

We would like to invite you to our 15th NLP meetup 🥳 This time it'll be all about LLMs with two exciting talks and an open discussion with you all. Let us start again gathering the excellent NLP community in Munich and having great topics, interesting discussions and fine networking 🚀

It has been almost a year since the launch of ChatGPT and what a year! At this Meetup, we will discuss two most pressing issues of GPT in industry: 1) battling hallucination 🔥 and 2) taming them in production with LLMOps 🔥

After the talks, we would like to have an open discussion with you all to reflect how and whether LLMs have changed your business 🤗🤗. We will share a link to gather your opinions beforehand to make the discussion easier.

=== Talks ===

Here are the talks and their brilliant speakers:

Hallucinations and Retrieval-Augmented Generation

• Speaker: Thomas Stadelmann, deepset
• Abstract: Hallucinations are one of the biggest challenges in running LLM-based applications as they can significantly undermine the trustworthiness of your application. Retrieval-augmented Generation (RAG) not only enables us to run LLMs on any data but also to mitigate hallucinations. However even with retrieval augmentation we cannot fully avoid them. In this talk Thomas will give a short into RAG and show you current approaches on how to systematically detect hallucinations, paving the way for automating this critical issue.
• Speaker's Bio: Thomas is Open Source Engineer at deepset working on Haystack, deepset's Open Source LLM Framework and deepset Cloud, their complementary commercial product. He originally started his passion for today's LLM-based solutions in Information Retrieval and Search applications. Given the development over the last years in those fields it's not surprising that he got hooked on the topic and so he's become passionate about anything that has to do with search, text processing in general and of course RAG.

Leveraging open-source LLMs for production

• Speaker: Andrey Cheptsov, founder of dstack
• Abstract: In this session, Andrey will discuss the importance of open-source LLMs, recent open-source releases, techniques to adopt open-source LLMs, cost considerations, and frameworks to use open-source LLMs.
• Speaker's Bio: Andrey is the founder of dstack, an open-source toolkit for orchestrating GPU workloads across clouds. Prior to this, Andrey worked at JetBrains with the PyCharm team.

=== Agenda ===

• 18:00: Doors open
• 18:15 – 18:30: Welcome and Introduction of our host appliedAI
• 18:30 – 19:00: Talk 1: Thomas Stadelmann, deepset
• 19:00 – 19:15: [Break] Networking and refreshment
• 19:15 – 19:45: Talk 2: Andrey Chepstov. dstack
• 19:45 – 20:30: Open Discussion: One-Year after ChatGPT - how has it changed you/your company
• 20:30 - 21:30: [Networking] Get-together with sponsored food and drinks

=== Sponsor ===

And eventually, we want to thank appliedAI for hosting this evening. If you don’t know appliedAI yet, you will learn more in a short introduction at the beginning.

Looking forward to meeting you!

Best Bernhard, Christoph, Kristian and Kun