Travelling into a long weekend and looking back to this week’s family reunion event aka #AWSSummit Berlin.

Travelling into a long weekend and looking back to this week’s family reunion event aka #AWSSummit Berlin. To be honest, I’m a little exhausted but energised at the same time after meeting so many customers, partners and colleagues. Always good to reconnect.

Kind of pointless to list every valuable session of the summit here as it was way too many, but I really loved to see customers on stage sharing their success stories and providing insights into the joined journey which lead them there.

Christopher Lennan and Andreas Hankel shared idealo internet GmbH’s cloud migration journey and how #ML and  #GenAI are impacting the company. It was great to see how — as they told it — the migration into the cloud laid the groundwork that let them bring their #GenAI workload to production. Super interesting use case. Spoiler: Jakob Panten and I will present the use case at the #AWSSummit Stockholm highlighting the Prompt Engineering aspect. Save the date: 4th June. Looking forward to this 😀.

Verivox GmbH showcased their MLOps modernization journey - from on-prem to AWS, and from spaghetti code to a neat setup with end-to-end responsibility, CI/CD, and Infrastructure as Code with CDK. While I couldn’t listen to Dr. Elisabeth Günther, Lucas Mattos Cruz and Julian Lang, as I have been presenting at the same time, I do know the story and I’m really happy that now others can benefit from the experiences the team made as well. Impressive.

Raj Pandey and David Surey delve into how ProSiebenSat.1 Media SE used AWS Copilot in their cloud migration journey. Super exiting to see them sharing the valuable experiences they made during their journey.

On my way to my own session I also had the opportunity to listen into Tobias Matern session with #dfl . Amazing:)

Philipp and I dived into how to build your own conversational agent using Hugging Face and AWS! 🚀 We showcased how to deploy and use Meta’s Llama 3 on AWS Inferentia2. AWS Inferentia2 are purpose-built chips for deep learning inference, offering an alternative to GPUs for open LLMs from Hugging Face — inference only, for the model architectures the Neuron SDK supports. I’m still flashed by how easy using AWS ML Accelerators through HuggingFace felt in our demo. Both days the room was packed. Kudos to the ones who couldn’t get a seat but joined the entire session standing in the back, you really showed patience.

Always a pleasure to listen to Gregor Hohpe and Dirk. They delivered the session “Advanced integration patterns and trade-offs for loosely coupled systems”,. Although integration architecture is not a new topic, they still get a lot of questions about it - because drawing a line between two boxes is easy, but assessing the integration options and their trade-offs requires a multi-dimensional consideration.

The list could go on and on, but actually I need to catch my flight. Cu next year - we will have a “Fischbrötchen” ;)

• AI
• AWS
• Events
• Architecture
• Community