Matei Zaharia (born 1984 or 1985) is a Romanian-Canadian computer scientist and educator. He is the creator of Apache Spark, a unified analytics engine for large-scale data processing, and a co-founder of Databricks, a company that commercializes cloud-based data and machine learning platforms. His research has focused on making distributed computing and machine learning systems more accessible and efficient.
In 2026, Forbes ranked Zaharia and Ion Stoica as the richest Romanians, with a net worth of $5 billion each.
Early Life and Education
Zaharia grew up in Toronto, Canada, and graduated from Jarvis Collegiate Institute before attending the University of Waterloo. As an undergraduate, he was part of the university's programming team that won a gold medal at the International Collegiate Programming Contest in 2005, placing fourth in the world and first in North America. During his undergraduate years, he contributed to the open-source game 0 A.D., particularly in water rendering physics, and also helped develop the Age of Mythology mod Norse Wars, which was later adapted into the Age of Empires III scenario Fort Wars.
He later pursued graduate studies in computer science, focusing on large-scale distributed systems. He was affiliated with the AMPLab at UC Berkeley, where his doctoral research addressed scalability limitations in existing data processing frameworks.
Creation of Apache Spark
While at UC Berkeley's AMPLab in 2009, Zaharia created Apache Spark as a faster alternative to MapReduce, the dominant large-scale data processing paradigm. Spark introduced in-memory computing and a more flexible execution model, enabling significant speedups for iterative and interactive workloads. His PhD research on large-scale computing earned the 2014 ACM Doctoral Dissertation Award.
Spark quickly gained adoption in both industry and academia, becoming one of the most widely used open-source frameworks for big data. Its influence extended to machine learning workflows by providing a unified platform for data processing, model development, and deployment tasks.
Career and Databricks
In 2013, Zaharia co-founded Databricks, a company built around Spark and related technologies. He serves as its chief technology officer, guiding the technical direction of the platform. Databricks has grown into a major player in cloud data management, offering services for artificial intelligence and analytics.
He joined the faculty of MIT in 2015, then moved to Stanford University in 2016 as an assistant professor of computer science. In 2019, he received the Presidential Early Career Award for Scientists and Engineers. That year, he also spearheaded MLflow, an open-source platform for managing machine learning lifecycles, which addresses challenges like experiment tracking, model reproducibility, and deployment.
In 2023, Zaharia joined the University of California, Berkeley as an associate professor, returning to the institution where he developed Spark.
Research Contributions and Recognition
Zaharia's research has spanned distributed computing, resource management, and systems for machine learning. Early in his career, he co-developed dominant resource fairness, a generalization of max-min fairness for multi-resource allocation, which influenced resource scheduling in cluster managers. His later work on MLflow and behavioral approaches to model selection has contributed to the operational side of machine learning systems.
For his foundational contributions to data and machine learning systems, he received the 2025 ACM Prize in Computing. He has also been recognized for his mentorship and influence in the open-source ecosystem, with Apache Spark serving as a cornerstone for many modern data engineering and deep learning pipelines.
Legacy and Impact
Zaharia's work has had a lasting impact on how organizations process and analyze data. Apache Spark's success helped popularize the idea of unified, high-level APIs for distributed computing, reducing the complexity of building scalable data applications. Through Databricks, he has also popularized the lakehouse architecture, which combines data storage and management with machine learning capabilities.
His contributions to AI infrastructure, particularly in enabling scalable training and serving, have supported the broader growth of deep learning applications. His academic leadership continues to shape research in cloud computing and data systems.