ACM prize crowns Matei Zaharia

A prize that usually catches rising stars in computing just went to the engineer behind a tool millions of data jobs already run on every day. On April 8, 2026, the Association for Computing Machinery named Matei Zaharia the winner of its ACM Prize in Computing. (acm.org) The award is for early-to-mid-career researchers whose work has had broad impact, and ACM said Zaharia was recognized for building data systems that made large-scale machine learning, analytics, and artificial intelligence practical. ACM also lists him as an associate professor at the University of California, Berkeley, and co-founder and chief technology officer of Databricks. (acm.org) (awards.acm.org) The core problem he worked on is simple to describe: one computer cannot chew through today’s giant datasets fast enough. So engineers split the work across many machines, then spend years fighting the mess of keeping those machines in sync. (awards.acm.org) Zaharia’s best-known answer was Apache Spark, which he started at the University of California, Berkeley AMPLab in 2009 and open-sourced in early 2010. Spark’s pitch was that many data tasks run much faster when more of the working data stays in memory instead of being repeatedly written out and read back from disk. (spark.apache.org) (awards.acm.org) That sounds technical, but the everyday version is a cook keeping ingredients on the counter instead of walking back to the pantry for every step. Apache Spark later grew into a large open-source project and moved to the Apache Software Foundation in 2013. (spark.apache.org 1) (spark.apache.org 2) Spark mattered because companies did not want one engine for data engineering, another for statistics, and a third for machine learning. The Spark project describes itself now as a single engine for data engineering, data science, and machine learning on one machine or a cluster. (spark.apache.org) ACM did not honor Zaharia for Spark alone. Its citation also points to Delta Lake and MLflow, two open-source projects aimed at the next bottlenecks teams hit after raw processing speed. (acm.org) (awards.acm.org) Delta Lake is about making large data stores more trustworthy, so teams do not build artificial intelligence on top of half-written or inconsistent tables. Databricks says its platform is built on a “lakehouse” design that combines data storage, governance, analytics, and artificial intelligence on one foundation. (databricks.com 1) (databricks.com 2) MLflow tackles a different headache: once a team trains a model, somebody has to track which data, code, and settings produced it. ACM says Zaharia created MLflow to handle experiment tracking, reproducibility, model management, and deployment across different machine learning frameworks. (awards.acm.org) That helps explain why this prize landed on a systems builder instead of a model inventor. The glamour in artificial intelligence still goes to chatbots and giant models, but the money and the daily work inside companies increasingly sit in the plumbing that gets data clean, moves it reliably, and turns experiments into production systems. (acm.org) (databricks.com) Zaharia’s career maps almost perfectly onto that shift. Databricks says he co-founded the company with six other Berkeley researchers, and the company now sells a unified platform for data, analytics, and artificial intelligence built on many of the same ideas that started in open-source software. (databricks.com 1) (databricks.com 2) So the headline is not just that one professor won one prize on April 8, 2026. It is that one of computing’s prestige awards just went to the layer between raw data and artificial intelligence, the part most users never see and most companies cannot function without. (acm.org) (awards.acm.org)