Building cost-effective homelabs for AI workloads

In this session, Johan van Amersfoort will share lessons learned of his 7-year journey to build a homelabs to run AI workloads. His own AI journey started in 2017 when the first VDI by Day, Compute by Night concept was developed. Since then, het build many different apps, platforms and used his lab to know more about AI, Machine Learning, Deep Learning, Models, Neural Networks, and things like computer vision. What worked? What didn’t? Do you always need to invest in GPUs? Can you build it cost-effective (spoiler-alert: YES!)? Do you actually need hardware or can you run it in the cloud? He will share interesting insights, show demos, and will share some tips and tricks to kickstart your own journey!

FOLLOW US
---------------------------------------------------
Twitter: /   / vmugbe  
LinkedIn: /   / belgian-vmug  
Discord:   / discord  

CHAPTERS
---------------------------------------------------
0:00 intro
0:55 Building Homelabs for AI workloads
4:08 What is AI?
10:55 How to create a AI workflow
19:33 Hardware for AI
35:51 How to start?
37:07 Conclusion
40:12 Q&A