How Databricks CEO Ali Ghodsi went from doing research in a college lab to cofounding the $38 billion big data startup — and what he's learned about leadership along the way

• CEO Ali Ghodsi cofounded $38 billion Databricks with University of California, Berkeley researchers.
• The most important groundwork for building company culture was a strong founding team, Ghodsi says.
• Databricks' next move will be growing its workforce. It's set to reach 3,000 employees this year.

When Swedish-Iranian computer scientist Ali Ghodsi arrived at the University of California, Berkeley for a one-year research opportunity, he had no intention of staying with the university or founding a startup.

But Ghodsi, one of Insider's 100 People Transforming Business in 2021, stuck around. He extended his time in the research program and, shortly after, cofounded the $38 billion big data startup Databricks with other UC Berkeley researchers, including executive chairman Ion Stoica and chief technologist Matei Zaharia.

When the team behind Databricks first started the company, they wanted to make a difference in the cloud industry, Ghodsi tells Insider. Databricks' core products started out as projects in the research lab. AI and cloud weren't big, but Databricks decided to bet on those technologies, along with the decision to build open source software, Ghodsi says.

The combination of these three became huge advantages for Databricks and eight years after its founding, those bets have seemingly paid off. Ghodsi holds the title of CEO and, in August 2021, Databricks raised $1.6 billion at a $38 billion valuation. The startup's valuation soared sixfold in less than two years.

Along the way, the Databricks founders learned lessons that helped set the culture of the company, one that's innovative and not afraid to make big bets despite others believing those bets are wrong, according to Ghodsi.

"People were really skeptical," Ghodsi tells Insider. "People think you have to do support and services. We believe you can build a different business model."

Databricks is now growing to take on big data competitors like Snowflake. The startup is on track to end this year with 3,000 employees (called "Bricksters"), up from 2,500 in October, with plans to keep hiring at an accelerated pace. While many tech companies hit the brakes on hiring during the pandemic, Databricks hit the gas on hiring engineers.

As the company prepares for the next stage, Ghodsi shares how he went from programming video games as a kid to CEO — and his advice on maintaining an innovative startup culture.

A lifelong programmer, Ghodsi traveled to California to research experimental data projects

Growing up in Sweden, Ghodsi says he started programming games as a child.

"Now, you need a big studio and artists to start a game," Ghodsi said. "In the late 80s, you could do this on your own. That's what got me started. I loved computers."

In college, the hobby grew into a consulting firm, writing and selling software to businesses. Ghodsi pursued a PhD in computer science and took a faculty job in Sweden doing research on data processing and computer systems. Ghodsi arrived at UC Berkeley in 2009 for a year-long program to research machine learning and data processing.

The program was so much fun, Ghodsi stayed for another year, and then another, finding each year more interesting than the previous one, he said. Eventually, Ghodsi decided to keep working at UC Berkeley, which he called the "perfect match."

Working at European universities, Ghodsi says he was often shut down when proposing out-of-the-box research ideas, but "UC Berkeley was different. They have a, 'You can do it all' attitude. You can start writing software and change the world and question assumptions."

Ghodsi went on to cofound Databricks out of a UC Berkeley research lab in 2013.

Databricks founders invest in company culture from the get-go

The biggest lessons Ghodsi learned from cofounding and leading Databricks weren't about how to grow a startup or sell products — they were about managing people.

Early on, the leaders realized the importance of a strong founding team to set the precedent for the type of workplace the company wants to be, according to Ghodsi.

"The first 20 people you hire are the culture of the company," Ghodsi said. "Who you bring in, the first 10 to 20 people will determine what kind of company you'll be for years."

Databricks' founding team was extremely innovative, Ghodsi says, with backgrounds in research and creating open source data project Spark. As a result, the company has an innovative culture that isn't afraid to experiment with and research new projects, and it has since built other cloud and AI projects. Databricks also relies on a team-oriented approach to these new ideas, as researchers collaborate with each other from all over the world, according to Ghodsi.

The founding team focuses on collaboration and research to determine the facts as they make business decisions, which has helped the company be "less political" than others in similar positions, Ghodsi says. Otherwise, as companies grow bigger, different departments within the company may get siloed and start fighting each other.

"In Databricks, we try to be truth-seeking," Ghodsi said. "We cherish and promote and take care of people who will focus on what's best for the company. Truth-seeking is what helps customers the most."