Welcome to our upcoming meetup at Crowe Foederer in Eindhoven! We have a lineup of experts ready to dive deep into the world of Agentic AI and Hybrid architectures.
Crowe Foederer
Crowe Foederer is a national accountancy and advisory firm with 740 employees and 9 offices located in Amsterdam, Den Bosch, Oostzaan, Lelystad, Bladel, Roermond, Venlo, Venray and its headquarters in Eindhoven. We are part of the global Crowe Global network, which operates in 150 countries. Our expertise spans seven service lines: audit, financial services, tax, HR services, corporate finance, advisory and technology. With this comprehensive range of services, we are able to fully support both SMEs and large organisations in all their challenges. Our strength lies in coaching and personal involvement, enabling us to help clients as their partner in progress to reach the next level.
Program
- 17:45 – 18:15 🍕 Food
- 18:15 – 18:25 🎤 Welcome
- 18:25 – 19:15 🎤 Bram Kostermans & Leon Gerritsen | Agentic AI for Data Analytics in Production: From Data Foundation to Intelligent Action
- 19:15 – 19:30 ⏸️ Break
- 19:30 – 20:15 🎤 Mahdi Massahi | Hybrid Intelligence: Balancing Deterministic Logic with Generative Power
- 20:00-21:00 🥤 Drinks
Bram Kostermans | Product Owner & AI Engineer at Crowe Foederer
Bram Kostermans is the Product Owner of CynthAI at Crowe Foederer, with a background in AI engineering and several years of experience designing and deploying machine-learning systems in production. He holds Master’s degrees in Operations Management & Logistics from Eindhoven University of Technology (TU/e) and in Accounting & Control from VU Amsterdam.
Before joining Crowe Foederer, Bram worked on computer vision, analytics, and machine-learning solutions across the high-tech, sports analytics, and industrial automation sectors, focusing on building reliable AI pipelines, operationalizing models, and translating technical capabilities into practical business value. At Crowe Foederer, he leads the CynthAI development team, working at the intersection of product strategy, AI engineering, and data infrastructure to architect agentic AI systems, enhance model reliability, and build scalable tooling that enables CynthAI to deliver trustworthy analytics in enterprise environments.
Leon Gerritsen | Lead Data Scientist at Crowe Foederer
Leon Gerritsen is Lead Data Scientist at Crowe Foederer, where he helps organizations design and build modern data and AI platforms that deliver measurable business impact. He brings a practical, hands-on approach to data engineering, machine learning, and operational AI, always focusing on solutions that are maintainable, scalable, and aligned with real business needs.
He started his Data Science career in 2017 after a short period teaching mathematics, and has since worked across medical, e-commerce, and financial domains. Throughout these projects, he has developed a strong focus on improving and automating business processes using data and AI. At Crowe Foederer, he leads the data engineering discipline, guides clients through architectural and strategic decisions, and helps teams move from proof-of-concept to production-ready systems.
🎤 Agentic AI for Data Analytics in Production: From Data Foundation to Intelligent Action
Agentic AI takes us beyond simple prompting toward systems that can reason, plan, and act. But making this work in production starts with strong data engineering. In this session, we’ll show how we built CynthAI, our agentic AI solution for enterprise analytics, and how its performance depends on a reliable data foundation created with Microsoft Fabric and our Common Data Format.
We’ll explore how CynthAI uses LangGraph, how we design agent workflows for unpredictable user questions, and what it takes to monitor, test, and maintain quality at scale. You’ll get a practical look at the combined data engineering and AI challenges of running agentic systems in real-world environments.
Whether you're exploring AI agents or modernizing your data platform, this talk offers a clear view of what it takes to operationalize Agentic AI, from clean data to intelligent action.
Mahdi Massahi | Machine Learning Engineer at Xebia
Mahdi is a Machine Learning Engineer Consultant at Xebia, with a background in Computer Science and a dual focus on technical consultancy and education. He leverages his academic foundation to analyse problems through the lens of computational complexity, providing unique structural insights and problem-solving skills for complex engineering challenges.
🎤 Hybrid Intelligence: Balancing Deterministic Logic with Generative Power
Production-grade AI applications require a clear separation of concerns between generative capabilities and deterministic execution. But making this work efficiently starts with knowing what not to ask the LLM. In this session, we’ll show how to optimise your architecture by offloading functional logic to external tools and utilising pre-computed context data rather than relying on inference.
We’ll explore real-world examples where we moved logic from the prompt to the codebase, how to design for predictability, and why it outperforms dynamic generation. You’ll get a practical look at the massive gains in testability and cost-efficiency that come from reducing the cognitive load on your model.
Whether you're building a new MVP or refactoring a costly application, this talk offers a clear view of how to operationalise your AI architecture for maximum reliability and performance.
