The PhD Student Who Made Big Data Fast — Matei Zaharia

Many famous tools start as a student project. Few change an entire industry.

As a PhD student at UC Berkeley, Matei Zaharia created Apache Spark, software that made analysing huge amounts of data much faster.

Spark became one of the most popular open source projects in the world and led to Databricks, now one of the most valuable startups on Earth.

A young coding star

Matei Zaharia was born in Romania and moved to Canada with his family.

He studied at the University of Waterloo. There he competed in the ACM International Collegiate Programming Contest, one of the toughest coding competitions in the world, reaching the World Finals in 2004 and 2005.

Matei Zaharia (right) at the 2004 ACM ICPC World Finals in Prague (Wikimedia Commons, credit needed)

Creating Spark

UC Berkeley, where Matei created Spark (Wikimedia Commons, credit needed)

Matei went to UC Berkeley for his PhD. At the time, a system called Hadoop was the standard for big data, but it was slow for many tasks.

In 2009, Matei started Spark. It kept data in memory, making some jobs up to 100 times faster. It was released as open source, and companies everywhere started using it.

His PhD work won the ACM Doctoral Dissertation Award.

Databricks

In 2013, Matei co-founded Databricks with Ali Ghodsi, Ion Stoica and other Berkeley colleagues. He became chief technology officer.

He also stayed a professor, first at Stanford and later back at UC Berkeley, working on data and AI systems.

A Databricks booth at a tech event (Wikimedia Commons, credit needed)

Where he is now

Today, Matei Zaharia is co-founder and CTO of Databricks and a professor at UC Berkeley. In 2026, he presented new AI tools on stage at the Databricks Data + AI Summit.

Why the story matters

Matei Zaharia's story shows how research can become real world impact. He solved a problem he saw in the lab, shared it openly, and helped millions of developers.

For young builders, the lesson is simple. Open source can be your launchpad. Solve a real problem and share the solution.

He did not keep his idea in a lab. He lit a spark for the world.

The Record

From: Romania; raised in Canada

Studied: University of Waterloo; PhD, UC Berkeley

Created: Apache Spark

Built: Databricks (co-founder and CTO)

Also: Professor, UC Berkeley

Timeline

2004 to 2005 Competes at the ICPC World Finals for Waterloo.

2009 Starts Apache Spark at UC Berkeley.

2013 Co-founds Databricks.

2014 Wins the ACM Doctoral Dissertation Award.

2016 Joins Stanford as a professor.

2026 Databricks valued at about $190 billion.