CNCF @ Piros

Hello everyone,

Let's start summer with a Bang. We are happy to announce or next meetup is going to be hosted by Piros, and we'll go deep into observability! ;)

What to expect:

1. 17:30 - 18:30 - Welcome and Pasta
2. 18:30 - 19:30 - This is an intervention: No-code observability with service mesh and tracing

• Description: Let’s start by a common statement that normally brings so much pain to the wall between engineering and whoever is calling the shots: “I want great observability without touching a line of code”. That’s simply not going to happen. You do get someobservability out of the box without touching your code base but, let us bring some good and bad news. The bad ones: it’s not going to be that great. The good: it’s a start! And, for what is worth, just reflect that when you are adding an inoffensive debug log to your code, you are indeed instrumenting your code. So, believe us, there’s nothing to fear!
  With this context in hand, given that instrumenting the code for observability is tedious, and distributed tracing heavily relies on context propagation it’s natural to wonder if the mesh can be configured to do all that “observability thingy” for us. The short answer is “no”, but that it can do a lot.
  In this talk we will show the path from zero to hero in observability with the service mesh and, how by adding some bits of instrumenting your code, you can finish reading your debugging book with a happy ending. And the best part… It’s Open Source!
• Speaker: Jose Gomez-Selles and Israel Blancas Alvarez

3. 19:30 - 19:45 - Break
4. 19:45 - 20:45 - Service mesh in practice: protecting (and building) AI models

• Description: During this session, we will explore a practical implementation of a service mesh to safeguard an AI model. We will discuss the benefits of using a service mesh in this context and providing high-level explanations of AI concepts to ensure accessibility for all attendees, regardless of their familiarity with AI. The session will then delve into AI/ML: implementing an MLOps workflow on Kubernetes using KubeFlow. Finally, we will demonstrate simpler methods for fine-tuning large language models (LLMs).
• Speaker: Maarten Vandeperre