Observability in LLMOps (different levels of scale) + other resources

July has been a productive month, and we have some exciting content to share! This issue covers insights gathered during the AI Engineer World’s Fair 2024, including the Aurimas Griciūnas' talk on observability in the LLMOps platform and pipeline, a step-by-step guide on how to build LLM applications with vector databases, and more.

Enjoy!

Case studies & practical MLOps

> 3 Takes on End-to-End For the MLOps Stack: Was It Worth It? - If you're facing a tough choice between selecting standalone components or an all-inclusive end-to-end platform for your MLOps stack, in this article, Ricard Borràs Navarra , Médéric HURIER , Maria Vechtomova help you decide by reflecting on their experience.

> Mikiko Bazeley: What I Learned Building the ML Platform at Mailchimp - Up next, we have an article by 👩🏻💻 Mikiko B. , outlining the learnings and challenges she has faced while on the ML platform team at Mailchimp, building infrastructure and setting up the environment for development and testing.

Guides & tutorials

 > Building LLM Applications With Vector Databases - Then, for those of you interested in vector databases, in this blog post Gabriel Gonçalves covers the essentials of iterating on efficient LLM applications, from Naive RAG to more complex topics like hybrid search strategies and contextual compression.

> Adversarial Machine Learning: Defense Strategies - Lastly, as ML models gain importance in key business applications, the potential for malicious attacks is also growing. Developing strong defenses becomes critical (especially in high-risk sectors). In this article, Michał Oleszak reviews common attack strategies and the latest defense mechanisms.

AI Engineer World’s Fair 2024

Now, it's time to switch to some video content!

Post-AI Engineer World’s Fair, we’ve launched a GenAI/LLM-themed interview series featuring Sandra Kublik , Jose Nicholas Francisco , David Li ☀️ , Francisco Ingham , Sam Julien , Alison Cossette , Jeronim Morina , YK Sugi , and more. They discuss everything from generative AI use cases and challenges to near-future predictions.

And if you haven’t seen it yet, at the event, Aurimas Griciūnas spoke about the importance of observability in the LLMOps platform and pipeline, focusing on areas that benefit from infrastructure scale. You can watch the full recording here.

That's it for our monthly update! If you find it valuable, please share it with your network.

Cheers!