# Patrick Wendell

> Patrick Wendell is a Databricks co-founder and vice president of engineering who helped build early versions of Apache Spark, established its three-month release cadence, and now leads major parts of Databricks' AI effort. His route from Princeton to Berkeley's AMPLab and then into a cloud company began with a practical surprise: users wanted the team's free software, but incumbent vendors did not want to distribute it.

- **Role:** Co-Founder and VP of Engineering at Databricks
- **Organizations:** Databricks, Apache Spark, UC Berkeley AMPLab, Princeton University
- **Education:** B.S.E. in Computer Science, summa cum laude, Princeton University, M.S. in Computer Science; began PhD study in the AMPLab, University of California, Berkeley
- **Known for:** Co-founded Databricks in 2013., Helped build initial versions of Apache Spark and became a founding committer and Project Management Committee member., Managed several Apache Spark releases and instituted its three-month release cadence.

## Career timeline

- **2011** — Joined UC Berkeley's AMPLab after Princeton to work with the distributed-computing group that became the founding team of Databricks.
- **2011-2013** — Helped build early versions of Apache Spark while conducting graduate research on optimizing large-scale analytics workloads.
- **2013** — Left the PhD track and co-founded Databricks with six other members of the Berkeley research community.
- **2014-2015** — Served as release manager for multiple Spark releases, maintained core-engine subsystems, and established a three-month release cadence.
- **2023** — Presented Databricks' Lakehouse AI capabilities at the Data + AI Summit engineering keynote.
- **2025-2026** — Led major Databricks AI and data-science product work, including Genie and Unity Gateway, while overseeing partnerships with leading model providers.

## Achievements

- Co-founded Databricks in 2013.
- Helped build initial versions of Apache Spark and became a founding committer and Project Management Committee member.
- Managed several Apache Spark releases and instituted its three-month release cadence.
- Graduated summa cum laude from Princeton with a B.S.E. in Computer Science.
- Earned an M.S. in Computer Science from UC Berkeley, researching optimization of large-scale analytics workloads.
- Leads Databricks AI efforts including Genie and Unity Gateway and partnerships with major model providers.

## Latest updates

- **2026-08** — Shared Databricks' analysis of internal AI coding costs, describing model choice, routing, adaptive budgets, and context optimization as ways to reduce unit costs by as much as 90% in some scenarios.
- **2026-06** — Helped present Databricks' expanded OpenAI partnership and discussed moving AI systems from research and demos into governed production use.
- **2026-06** — Discussed Unity AI Gateway controls intended to prevent runaway enterprise AI spending and improve visibility into model costs.
- **2025-11** — Appeared on The Economic Times' ET in the Valley podcast to discuss Databricks' early cloud bet, AI economics, and engineering talent in India.

## Links

- LinkedIn: https://www.linkedin.com/in/patrick-wendell
- Twitter/X: https://x.com/pwendell
- GitHub: https://github.com/pwendell
- YouTube: https://www.youtube.com/watch?v=wavBmKNBeRA

---

Profile page: https://yespress.io/patrick-wendell
Published by YesPress — https://yespress.io
Last updated: 2026-09-26