Conflux Connect - The Future of IT

Save the date!
Date: 27 August 2026 - 17:00 – 21:00
Location: Europalaan 93, Utrecht

Artificial Intelligence is rapidly becoming part of every developer's toolkit. From AI-powered development environments and agent-based applications to token optimization and running local Large Language Models, the landscape is evolving at an unprecedented pace.
During this Conflux Connect meetup, we will explore the technologies, patterns, and practical approaches that are shaping modern AI-driven software development. Whether you're already building AI-enabled applications or simply curious about where the industry is heading, this evening will offer valuable insights and real-world experiences.

Topics include:

• Native AI development
• AI harnesses and developer tooling
• Working with tokens and context management
• Local and self-hosted LLMs
• Practical applications and lessons learned

This event is intended for software developers, architects, cloud engineers, data professionals, and anyone interested in the future of software development.

Program:
17:00 – 18:00 Walk-in + Dinner
18:00 – 18:30 Agentic GitOps in Practice – by Bobby Prins (Conflux)
18:30 – 18:40 Break
18:40 – 19:10 (Im)possibilities of using local models for AI assisted coding – by Arie Weeren (BigData Republic)
19:10 – 19:20 Break
19:20 – 19:50 Vibe Coding: The Good, the Bad & the Ugly – by Gareth Thomas
19:50 – 21:00 Networking Drinks

About the talks:
Agentic GitOps in Practice
Bobby Prins – Conflux
What happens when an AI agent helps manage a Kubernetes cluster? In this session, Bobby shares his experience managing a homelab using Flux and Hermes to explore the combination of GitOps workflows and AI-driven automation. He will discuss what worked, what didn't, and address questions such as how much control should be handed over to AI and what the security implications are of running autonomous agents.

(Im)possibilities of using local models for AI assisted coding
Arie Weeren – BigData Republic
Can consumer hardware really handle local AI coding, or is it still more of a novelty? Arie explores the practical reality of running local Large Language Models on a variety of developer machines, from MacBook M3 Pro systems to PCs equipped with 16GB VRAM and high-memory configurations.
Drawing on real-world experience from complex Rust and Python projects, he will demonstrate what is actually possible, where the limitations lie, and how hybrid workflows combining local models with services such as Claude can support modern software development.

Vibe Coding: The Good, the Bad & the Ugly
Gareth Thomas
AI-powered "vibe coding" is changing the way software is built, making it possible to go from idea to working prototype in record time. But what happens after the initial excitement?
Using real-world examples from the development of Meetello.events with GitHub Copilot and Claude, Gareth explores where AI genuinely accelerates development, where it starts to break down, and the challenges that emerge as projects grow. Topics include technical debt, debugging AI-generated code, maintaining software architecture, testing, and practical workflows that combine AI with sound software engineering principles.
Whether you're an experienced developer, a hobbyist, or simply curious about AI, this session offers a realistic perspective on what AI-assisted development can—and cannot—do in production environments.