Munich Datageeks July Edition

As always, there will be plenty of time for networking, discussions, and snacks & drinks courtesy of our host.

📍 Event Details
📅 Date: Thursday, July 30th, 2026
📍 Venue: Celonis Munich
📫 Address: Theresienstraße 6, 80333 München

🎤 Perspective Matters! Automated Performance Analysis of Interacting Business Objects
by Jan Niklas van Detten, Celonis

Enterprise information systems track the interactions between business objects of various types, such as sales orders, items, and deliveries. Object-centric process mining aims to analyze these records to ideally improve business-relevant performance measures, such as the cycle time. However, this analysis requires a clear notion of which objects belong to a joint execution of the processes, commonly called a case. Each such distinct case notion intuitively corresponds to one of many analytical perspectives on the underlying business process. Technically, these case notions define which interactions are considered relevant for any given analysis goal. This poses a significant challenge in practice due to a lack of automated methods and the exponential number of possible case notions to choose from. In this presentation, we will discuss methods to automatically align the case notion selection with any given analysis goal, such as the automated performance analysis.

About Jan Niklas van Detten:
Jan Niklas van Detten is currently a PhD student at RWTH Aachen University in association with the Process Mining Technology Department at Celonis and has previous startup experience in data management and analytics. His current research focuses on the automated analysis of interacting business processes based on event data. In this talk, he will present patented state-of-the-art methods for the automated performance analysis in such processes.

🎤 Searching lost pets using neural networks
by Olga Trofimova

Every year, thousands of pets go missing, leaving owners relying on posters, social media, and chance. But what if finding a lost pet worked like face recognition on your phone?
In this talk Olga shares how modern computer vision and deep learning techniques can be applied to identify animals by their faces. She will describe a process of creating a real-world system that detects and recognizes cats and dogs using neural networks, enabling automated matching between lost and found pets.
The talk walks through the full pipeline: from training detection and recognition models, to overcoming one of the biggest challenges in AI -- data. Learn the basics of recognition pipelines and how noisy, unstructured pet images from the internet, shelters, and crowdsourcing were transformed into a large-scale dataset.
This project shows how AI can move beyond theory to solve an emotional, real-life problem helping reunite pets with their owners.

About Olga Trofimova:
Olga is a Computer Vision Engineer and former member of the Face Recognition Team at NTechLab, a global leader in AI-powered video analytics. She specializes in biometric recognition systems and deep learning.