· de makersfabriek, zwolle, overijssel
Meetup No. 3
geweest- Marin Niehuesai-is-easy-until-you-add-people-inside-the-chaos-of-implementing-aiover deze talk
Het verhaal van binnenuit over AI invoeren bij een wereldwijde fabrikant: legacy-systemen, ontbrekende data, sceptische engineers en een organisatie die niet op de verandering is ingericht.
AI is supposed to be the future. At least, that's what all the PowerPoint decks say.
The pitch always sounds great - smarter production lines, real-time process optimization, predictive everything.
And then you actually try to make it happen.
It turns out that AI isn't the hard part - everything else is. The biggest blockers? Legacy systems that refuse to cooperate, missing data that no one realized was missing, skeptical engineers who don't trust "some algorithm," and processes designed for a world that existed long before AI.
But most of all, it’s the people and the organization. Most companies like the idea of AI far more than the change it actually requires
Roles, responsibilities, decision paths -none of it fits. And no one wants to be the first to do things differently.
This is the inside story of embedding AI in a global manufacturer, a project I led from start to finish. The goal was ambitious, the potential was huge, but reality had other plans. I'll walk you through what actually happened-the unexpected resistance, the messy pivots, and the moments that nearly killed the project.
What you'll walk away with...
... a real-world AI case study-not the polished version, but the unfiltered reality.
... why AI isn't just about technology - it's about people, processes, and organizational politics.
... how to deal with resistance, broken data, and invisible roadblocks.
... lessons from the trenches. What works, what doesn't, and what to avoid.
... how to avoid AI mistakes before they derail your next big thing.
programma - Arjen Wiersma7-rules-for-secure-coding-with-aiover deze talk
Echte voorbeelden van hoe een simpele prompt aan een AI-assistent een beveiligingslek kan openen, en zeven regels om Copilot en Cursor veilig te gebruiken.
Using AI assistants like GitHub Copilot and Cursor feels great, you can "vibe" with the AI and get working code, fast. But this speed, if you're not careful, can create serious security problems. This talk goes beyond vague warnings and shows you real-world examples of how a simple prompt for an api call or an error message can accidentally create a major security hole for attackers.
But this session isn't just about the problems; it’s about the solutions. You will leave with "7 Rules for Secure Coding with AI," a straightforward guide you can start using right away. You'll learn practical skills to use AI safely and effectively.
programma - Ayodeji Ogundareopen-source-and-the-new-ai-stack-where-developers-fit-inover deze talk
Een kaart van de open source AI-stack, van modellen en infrastructuur tot tooling en developer-interfaces, en waar developers kunnen bijdragen zonder ML-kennis.
The AI world is moving fast, but it is also leaving many developers behind. Between foundation models, inference engines, agent frameworks, and prompt orchestration tools, it’s hard to know where to plug in, especially if you’re not training models yourself.
Here’s the good news: open source is powering the new AI stack, and developers of all kinds are needed to help shape it.
In this talk, we’ll break down the modern open source AI stack into its core layers, from models and infra to tooling and developer interfaces, and show how developers can contribute meaningfully at every level. Whether you're a backend dev building APIs, a frontend engineer designing UX for agents, a DevOps engineer setting up inference infra, or someone improving docs and onboarding, there’s a place for you.
This talk is for anyone who wants to transition from curious observer to active contributor in the open source AI movement.
Key Takeaways:
- A clear, visual map of the open source AI stack
- Specific OSS projects and roles developers can contribute to (even without ML expertise)
- What great DevX looks like in open source AI and how to improve it
- How to get started: contribution paths, communities, and first PR tips
programma
Programma
$ devpulse meetup --no 3
do 4 sep 2025 · de makersfabriek, zwolle · 3 talks · deze avond is geweest
Drie talks over AI in de praktijk: AI invoeren bij een wereldwijde fabrikant, veilig programmeren met AI-assistenten, en waar developers passen in de open source AI-stack.
- deuren opende makersfabriek, zwolle
- eten
- talkai-is-easy-until-you-add-people-inside-the-chaos-of-implementing-ai --speaker "Marin Niehues"
- talk7-rules-for-secure-coding-with-ai --speaker "Arjen Wiersma"
- talkopen-source-and-the-new-ai-stack-where-developers-fit-in --speaker "Ayodeji Ogundare"
- borrel
De talks.
· talk
AI Is Easy - until you add people: Inside the Chaos of Implementing AI
Marin Niehues
Het verhaal van binnenuit over AI invoeren bij een wereldwijde fabrikant: legacy-systemen, ontbrekende data, sceptische engineers en een organisatie die niet op de verandering is ingericht.
AI is supposed to be the future. At least, that's what all the PowerPoint decks say.
The pitch always sounds great - smarter production lines, real-time process optimization, predictive everything.
And then you actually try to make it happen.
It turns out that AI isn't the hard part - everything else is. The biggest blockers? Legacy systems that refuse to cooperate, missing data that no one realized was missing, skeptical engineers who don't trust "some algorithm," and processes designed for a world that existed long before AI.
But most of all, it’s the people and the organization. Most companies like the idea of AI far more than the change it actually requires
Roles, responsibilities, decision paths -none of it fits. And no one wants to be the first to do things differently.
This is the inside story of embedding AI in a global manufacturer, a project I led from start to finish. The goal was ambitious, the potential was huge, but reality had other plans. I'll walk you through what actually happened-the unexpected resistance, the messy pivots, and the moments that nearly killed the project.
What you'll walk away with...
... a real-world AI case study-not the polished version, but the unfiltered reality.
... why AI isn't just about technology - it's about people, processes, and organizational politics.
... how to deal with resistance, broken data, and invisible roadblocks.
... lessons from the trenches. What works, what doesn't, and what to avoid.
... how to avoid AI mistakes before they derail your next big thing.
· talk
7 Rules for Secure Coding with AI
Arjen Wiersma · Cyber Security Consultant at Scyon
Echte voorbeelden van hoe een simpele prompt aan een AI-assistent een beveiligingslek kan openen, en zeven regels om Copilot en Cursor veilig te gebruiken.
Using AI assistants like GitHub Copilot and Cursor feels great, you can "vibe" with the AI and get working code, fast. But this speed, if you're not careful, can create serious security problems. This talk goes beyond vague warnings and shows you real-world examples of how a simple prompt for an api call or an error message can accidentally create a major security hole for attackers.
But this session isn't just about the problems; it’s about the solutions. You will leave with "7 Rules for Secure Coding with AI," a straightforward guide you can start using right away. You'll learn practical skills to use AI safely and effectively.
· talk
Open Source and the new AI Stack: Where Developers fit in
Ayodeji Ogundare · Developer Advocate at Adyen
Een kaart van de open source AI-stack, van modellen en infrastructuur tot tooling en developer-interfaces, en waar developers kunnen bijdragen zonder ML-kennis.
The AI world is moving fast, but it is also leaving many developers behind. Between foundation models, inference engines, agent frameworks, and prompt orchestration tools, it’s hard to know where to plug in, especially if you’re not training models yourself.
Here’s the good news: open source is powering the new AI stack, and developers of all kinds are needed to help shape it.
In this talk, we’ll break down the modern open source AI stack into its core layers, from models and infra to tooling and developer interfaces, and show how developers can contribute meaningfully at every level. Whether you're a backend dev building APIs, a frontend engineer designing UX for agents, a DevOps engineer setting up inference infra, or someone improving docs and onboarding, there’s a place for you.
This talk is for anyone who wants to transition from curious observer to active contributor in the open source AI movement.
Key Takeaways:
- A clear, visual map of the open source AI stack
- Specific OSS projects and roles developers can contribute to (even without ML expertise)
- What great DevX looks like in open source AI and how to improve it
- How to get started: contribution paths, communities, and first PR tips
Waar.
- locatie
- De MakersfabriekCeintuurbaan 15, gebouw ZF (Het Ketelhuis)Zwolle, Overijssel
- route
- Parkeer op het terrein van De Makersfabriek: rij het oude Philips-terrein op bij de McDonald’s. Loop naar gebouw ZF (Het Ketelhuis), voorbij ZWINC, en sla linksaf bij de ijsbeer. De ingang is aan de achterkant van het gebouw.routebeschrijving
Spreek op DevPulse
$ pwd /devpulse
cd ./speakEen talk over iets dat je bouwde, brak of leerde. Stuur een titel en twee regels.[email protected]