Maria Vechtomova, Marvelous MLOps, and Sylvain Kalache, Rootly
The LLM narrative focuses on how AI empowers developers to write code faster, but far less attention is given to its impact on SREs and platform engineers.
This session will explore how AI is reshaping the SDLC after the code is written: from CI/CD pipelines, deployments, scaling, and monitoring, to incident management, reliability tooling, and emerging disciplines like LLMOps. We’ll discuss what's genuinely new, what remains unchanged, and what might be more hype than substance.
Questions we could explore:
- As developers write more code, faster, how does this impact CI/CD pipelines?
- Can deployment velocity match the new development rhythm?
- ML applied to monitoring isn’t new; are LLMs bringing new capabilities?
- How will these shifts affect incident responders?
Maria Vechtomova is in the Data and AI for almost 12 years, focusing most of her career on MLOps. Maria believes that everyone should learn MLOps, as machine learning models only start living once in production. She is a co-founder of Marvelous MLOps, teaching courses on Maven: MLOPs and LLMOps with Databricks, and writing a book for O'Reilly.
Sylvain leads AI Labs at Rootly, an initiative that aims to augment reliability engineering in the era of AI. Under his leadership, the lab has developed open-source prototypes, tools, and research collaborations, sponsored by organizations like Anthropic, Google DeepMind, and Google Cloud.
author = {Maria Vechtomova and Sylvain Kalache},
title = {Tea, Pipelines, and Retries: A Practical Guide to {MLOps}},
year = {2025},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}