Ion Stoica is a professor of computer science at UC Berkeley, Co-Founder and Executive Chairman of Databricks, and a key architect of the Apache Spark project. Most recently, he’s the Co-Founder of Anyscale, which leverages the open source Ray framework developed in-lab to enable scalable AI workloads, much like Spark revolutionized large-scale data processing.

In this episode of The Data T, we chat with Stoica about his illustrious career, how his obsession with solving hard technical problems led him from networking research to peer-to-peer video, Apache Spark, and ultimately Databricks. He recounts turning Spark’s open-source momentum into a successful enterprise business, crediting speed of execution and targeted hiring for the company’s rise and urging founders to move fast and recruit experienced operators early. Stoica warns that tomorrow’s workloads will demand vertically integrated, multi-accelerator systems. Optimistic yet realistic about AI, he sees reliability and “human-in-the-loop” workflows as today’s gating factors and advises data professionals to embrace continuous learning as the industry accelerates.

Hosted by Armon Petrossian and Satish Jayanthi, co-founders of Coalesce.

Key topics:

• The origins of Apache Spark and Databricks
• Commercializing open source projects
• Scaling AI infrastructure complexity
• Advice for data practitioners

Resources:

• About Coalesce: https://coalesce.io/about/
• Coalesce podcast archive (The Data T): https://coalesce.io/podcast/