Skip to main content

Ajay Gulati

Co-Founder and CTO, ZeroStack

Since October 2008, Rice University CS Ph.D. alumnus Ajay Gulati, co-founder and CTO at ZeroStack, has been granted 36 U.S. Patents and has submitted applications for 34 others. Clearly, he is a researcher who understands how to define and approach a problem.

“I really think programming or software development is just a means to an end,” he said. “At the end of the day, you have to solve a real problem. You can build a software app or you can build a computer system on which others build their applications.”

Gulati said his Ph.D. research focused on storage performance virtualization, research that could later be applied to real systems such as storage systems or network routers. He said, “For me, systems presented the most interesting challenges. If you solve a problem here, then it benefits a lot of other people who will build on top of it.  I wanted to increase capacity on a single storage system so that multiple developers could work in tandem without impacting each other.”

When Gulati graduated from Rice, he had no plans to launch his own company. Instead, he went to work as a VMware senior staff engineer in resource management research and development, specializing in storage systems and scheduling algorithms.

“I spent seven years in industry and found a compelling problem that needed to be solved. I kept working on it until I and a friend decided to launch our company. I left VMware in August 2013 and the startup, ZeroStack, launched nine months later,” said Gulati.

The two friends started out with a small working group, talking to potential customers, doing market research, and building an initial prototype of what the solution would look like.

Gulati said, “When we reached the point where we could show how we were solving the problem, provide a schedule for release, and a list of customers, then we started reaching out to venture capitalists. We ran without VC funding for about six or eight months. When we got VC funding, then we started building.”

His Ph.D. training helped prepare Gulati for the surprises he would encounter building a company from the ground up.  “Completing a Ph.D. provides you with confidence that you can solve a problem even if it has not been done before. You really need that confidence as an entrepreneur. You are going to build a company around solving problems that you’ve never solved –  maybe no one has ever solved before – and you need confidence in yourself to take that risk.”

He said the willingness to take that risk is the difference between entrepreneurs, who are diving into the unknown, and software engineers, who are working on projects according to someone else’s specifications.

“When I was working on my Ph.D., I was focused on the research and pursuing knowledge, not really thinking about entrepreneurship,” said Gulati. “It was only later that I realized the Ph.D. program gave me the confidence to be able to face the unknowns when building something new.”

Because he was working closely with his partner and the other startup team members, Gulati discovered the unknowns and complexities of bringing ZeroStack to market extended beyond technology problems.

“Once the product is ready, now you have to work on another set of problems, which while not technical, are equally important,” he said.

“What should be the position of the product when you launch? How will you talk about it in sales pitches and on your website? Often engineers are perceived as doing the heavy lifting, but the other side – marketing, customer development, sales – those areas have their own challenges. Nothing is easier or harder in one area than any other part of the company.”

Gulati said a marketing specialist has to understand the company’s vision and the product being built, but the product manager and engineers also receive valuable input from the marketing person. “When you hire a person to figure out how to best position the product, they also have questions and you have to work with them to get that product out the first time,” he said.

Based on his experience with ZeroStack, Gulati now has a clearer understanding of how the product manager and marketing specialist work together. He said, “The product manager talks to customers to figure out what makes a complete product. You can’t put 100 features on the first release, probably only 10, so the product manager has to figure out which 10 features make a viable product.”

“The product manager defines the project, works with engineers on the build, and can talk to both sides – the customers and the engineers. The marketing person then takes the project out to the market. That is where the marketing specialist needs to understand the landscape, competitors, and what is unique about our product. And you can’t build a business on a “me, too” product. Everything has to have a differentiator.”

The differentiator for a ZeroStack private cloud is its self-management. Gulati said other vendors sell private clouds but the ZeroStack product is different because it is self-driving and self-managing. “Our customers don’t need a deep team to build and manage our software. Back to the marketing perspective, our tag line tells both what we are selling and how it is different than others. The private cloud is what we’re selling and self-driving is what makes it different.”

Gulati said he had to learn the business side of their company quickly. “I feel like I’ve gone through an MBA class, learning as I go,” he said. “But it wasn’t a class, it was real. If you make a mistake, you have to recover from that mistake.”

Although Gulati is thriving along with ZeroStack, he misses the satisfaction he found in deep research and publishing. “I used to participate in conferences and committees and I’m still interested in research and solving interesting problems, but I don’t have time to publish papers anymore. That is the biggest thing that’s changed from working for an established company to starting our own.”

“When you are running as fast as I am right now, launching a startup, there’s just no time to spend four to six weeks working on a paper. But we’ve developed a lot of solutions, good solutions. When things slow down, I’d like to start publishing papers again.”


This article was originally written by Carlyn Chatfield on the Rice University Department of Computer Science CS Profiles website.