AI for Systems

In the last decade, data management has been steadily moving to-wards public clouds, turning DBMS "builders" into "operators". The combined scale and vertical integration gives us the opportunity to optimize systems holistically using a telemetry-based feedback loop. These events aligned with fast advances in the field of Data Science to create the perfect conditions for an explosion of "ML for Systems" solutions. More recently, massive improvements in Large Language Models (LLMs) has introduced another intriguing element into the mix. In this panel, we explore questions around the practical adoption and robustness of ML for Systems solutions, as well as the potential role for LLMs, spanning very concrete engineering considerations and very open-ended "future of this industry" type thinking.