AI at IBM

When:
June 13th, 2024
Location:
IBM Client Innovation Center Netherlands
Gedempte Zuiderdiep 12,
9711 HG Groningen

*** Please be aware ***
Security policy at IBM requires us to supply a guest list to them with your full legal name (and company name, if available). As organizers of the meetup, we will try and reach out to you beforehand to verify your name and company. We will see to it that the list gets deleted after the meetup.
*************************

Agenda:
17:30 - walk-in and food
18:15 - Welcome by GroningenML and IBM
18:30 - AI Hub Noord
18:45 - First presentation - Building a chatbot for Gemeente of Groningen using RAG
Speaker: Remco Pals – Machine Learning Engineer @IBM CIC
19:15 - short break
19:30 - Second presentation - How MLOps can help the future you
Speaker: Thomas Brinkman – Data Architect + Co-founder @Datadenkers
20:00 - Third presentation - Using WatsonX to enhance credit risk analysis
Speaker: Filippos Pappas – Data Scientist @IBM CIC
20:30 - Closing statements by IBM/GroningenML
20:35 - Drinks and networking
21:10 - Meetup ending

AI Hub Noord
AI Hub Noord-Nederland plays a crucial role in advancing AI innovation, collaboration, and awareness in the region. They are the linking pin to the National Network of AI and part of the AINED program. In this talk, Bart will explain more about the hub's activities and how he might be able to help you.

Building a chatbot for Gemeente of Groningen using RAG
At IBM, we created a RAG-based chatbot for Gemeente of Groningen, leveraging the open data, all in Dutch language. Additionally, we implemented a variety of ways to improve both the retrieval and the generation of the framework, which will both be explained in theory and shown in a short, interactive demo.
Keywords: LLMs, VectorDBs, WatsonX, llama3, RAG, HyDE, GPT4, streamlit

How MLOps can help the future you
Why and how to implement several components of the MLOps stack so that it supports traceability and to make your future life easier.

Using WatsonX to enhance credit risk analysis
Financial institutions face challenges in identifying and managing financial risks from delinquent clients. This lack of timely recognition of bankruptcy signals contribute to significant losses for their business. Developing tools that are leveraging powerful LLMs, can contribute to more fast, accurate and effective credit risk analysis process using internal and external data to provide justifications on the alerted flags.
Keywords: LLMs, WatsonX, llama3, PostgreSQL, PowerBI