Manasi Vartak '10 founded Verta to help companies build AI-enabled products faster than ever.
Manasi Vartak '10, founder of Verta
Manasi Vartak ’10 doesn’t shy away from the unknown. Now CEO and founder of AI start-up Verta, she came to WPI intent on discovery. With the drive to study the science and math she loves and the fearlessness needed to relocate her life from India to New England, Vartak knew her bold approach was necessary to have a big impact on the world.
As a high school student in Pune, India, Vartak knew pursuing her education overseas would allow her a particular freedom she craved. “If you go to college in India, there are more rigid disciplines,” she says. “If you are going to be an engineer, you study only engineering. If you want to go into biology, you study only biology or math. That’s fine if you want to specialize in only one thing, but I had interests in a lot of areas. To put them into buckets seemed very limiting. I wanted to be able to connect dots between disciplines, and be able to shape my own path.”
Always a builder, Vartak wanted an engineering college and a technical school so WPI made her first cut. But the university’s distinctive project-based learning approach especially intrigued her, she says. “I appreciated WPI’s hands-on nature,” she says. “I appreciated the IQP (Interactive Qualifying Project), the MQP (Major Qualifying Project). I wanted a place that was hands-on and where I had the freedom to explore. It taught me to stay interested, engaged, and constantly continue to learn.”
Vartak first saw the WPI campus in person when she arrived for New Student Orientation in 2006. Coming from a city of about 7 million people to Worcester was a big transition. Meeting students from all over the globe, and even those from different cultures and traditions within the United States, taught her to consider new, unfamiliar perspectives, an important skill for any modern business leader. She dove in to explore everything—academics, social life, and extracurricular activities.
“College makes you who you are as a person,” she says. “It teaches you grit and how to work through challenges because you are on your own. For me, WPI was all that. I learned a ton. I did so much, and my four years there were a blur. It was a really good experience for me.”
Vartak says she’s grateful to the alumni who paved the way and created generous scholarships that sealed her decision. “They made it possible for me to come to WPI,” she says. “I am extremely grateful for them and the school as a whole.”
As with all WPI students, humanities and the arts were part of Vartak’s foundation. “WPI really focuses on rounding you out as an individual. I got to take art classes, which, even though they don’t have anything to do with AI or machine learning, were really important because they helped shape who I became and how I look at the world.” she says. “And my writing classes especially have served me really well. As an entrepreneur and business leader, your communication skills have to be razor sharp. WPI helped give me that.”
Finding Her Footing
And the freedom she so craved? At WPI, it actually changed the course of her studies. “I don’t think I ever would have imagined studying computer science. Initially, I think physics was my intended major.” she says.
But a Scheme course and then one on Java helped her envision the enabling power of technology, and that opened up possibilities she had never considered. “I could work in a variety of disciplines and on a variety of problems and make an impact that was pretty wide,” she says. “I came to see computer science as a powerful tool to innovate and build with, and I gravitated toward that.”
Vartak credits Elke Rundensteiner, professor of computer science and founding director of the Data Science Program, with giving her the motivation and inspiration to take the direction she did. “She is the reason I even pursued a PhD,” says Vartak. “I got started in databases in her lab.” As a formative mentor, Rundensteiner introduced Vartak to research opportunities that were intellectually exciting.
“I came to see computer science as a powerful tool to innovate and build with, and I gravitated toward that.”
Close collaboration with Rundensteiner allowed Vartak to dive headlong into her newfound love of computer science. “At WPI, I got to explore a ton of things and see what resonated,” she says. “Sometimes the larger schools don’t let you do that. I had access to faculty and had opportunities I might not have had at larger schools.”
She says Rundensteiner’s ability to inspire her students by modeling curiosity and tenacity was, and remains, impressive. “She is amazing,” says Vartak. “She’s encouraging and will push you to do better. It’s a delicate balance. She assigns problems that are a good fit for the student but outside their reach. She makes sure you work with good grad students and have a good support system.”
Vartak’s computer science degree and her love of data shaped her journey and led her to a PhD program at MIT, where the seeds of Verta were planted. As she progressed through professional roles at tech giants including Twitter, Google, and Facebook, she developed as a leader. But the love of discovery and exploration cultivated at WPI continued to mature in her.
Creating a Valuable Product
In 2018, Vartak decided to capitalize on her years of experience and exploration and launch Verta, a new kind of platform that enables businesses who see value in data and AI to push boundaries. Just two years later, Verta secured a $10 million Series A round of funding from investors, a sign of the company’s potential.
She describes Verta as the first system of its kind, a solution to the decades-old divide between research and creating a product from that research. In concept, it is similar to a pharmaceutical process, she says, where scientists might find a useful chemical compound for a drug, but then it needs to become a useful, consumer-friendly product.
Data scientists and machine learning teams often struggle with similar issues. “A lot of data science is building something that is one-off,” she says. “Data scientists experiment a lot to arrive at algorithms—or models—that offer accurate, useful insights. Once you have a useful model, you have to figure out how to move the model into production—how do you scale it up, optimize it, and make it robust? The process is very different from what is required to build the models in the first place.”
Before Verta, the process left a gaping hole between current research findings and old data. AI is only as good as your data, says Vartak, so she wanted to match the most current data with ever-evolving models.
“And we also want to enable teams to evaluate how well a model is performing in the real world and relay that back to the researcher,” she says. “Ideally, you create a loop of information that links real-time results to the model the data scientists developed. If I have a great idea, I want to instantaneously be able to deploy it, and then learn from those relays and signals in order to improve models immediately.”
As she approached the beginnings of an innovative path, Vartak says she knew she had the determination and the skillset to deliver something no one else ever had. WPI projects, including work on ASSISTments with Neil Heffernan, director of the Learning Sciences & Technologies Program, taught her how to work fast and to experiment quickly, even with limited resources and impossible deadlines.
As with any kind of experimental journey, iteration is essential to success. “In today’s business context, market dynamics move extremely quickly—and machine learning models need to be able to keep up with data patterns. So if you can speed up the time it takes to go from experimentation to the next iteration, you can increase the speed of innovation.”
But as a new discipline, best practices for operationalizing machine learning are still being established. Vartak created Verta to help streamline these processes, allowing businesses across industries to extract more value from their data.
“If I have a great idea, I want to instantaneously be able to deploy it, and then learn from those relays and signals in order to improve models immediately.”
The uncertainty of starting a company from nothing is invigorating to Vartak. “Start-ups are full of uncertainty,” she says. “They are roller coasters. Every day is a new challenge, but every day is a new opportunity. You work very hard and if you do everything right, you’ll have a successful business.”
Vartak says learning from mistakes is how start-ups—and start-up leaders—become better. “You can fail and when you fail, you can get over it and can pick yourself up and keep going,” she says. “All the small experiences you have along the way help you accomplish big things.”
Being able to understand the customer’s need is essential. “There’s such a big gap between ‘I have an idea’ or ‘I have a solution’ and ‘I am going to build it and deliver it and make sure it works for the customer,’” she says. “I’ve come to appreciate what it takes to build a world-class product. That’s the trial and error and the unique path to a start-up.”
As Vartak continued with her intense doctoral path and the launch of Verta, she kept Rundensteiner’s example in mind. “She’s an exceptional researcher and has a great family,” she says. “She has a good sense of humor and is a well-rounded individual. With a young family myself, I am realizing the constant trade-off and am appreciating women like her even more.”
She also realizes the need for role models. “There are few tech women leading companies, and it’s not for lack of talent,” Vartak says. “There are a lot of reasons why I started Verta. But one of them was to show other women, ‘This is a possible career path for you, and you can do it well.’”
Those experiences have even shaped how Vartak runs Verta. Knowing she doesn’t mirror many traditional ideas of tech startup leaders, she is consciously developing a diverse team at the software company. Representation of different experiences, backgrounds, ages, and genders, to name a few, matters to retention, to job satisfaction, and to a global workforce.
A Data-Based Future
“At every stage, the challenges are different,” she says. “Building something like Verta takes effort, patience, and perseverance, and I learned a lot of those at WPI. Professional satisfaction depends on being able to identify the problems you care about solving. That’s what make work fun, and that has to be the subtext.”
For Vartak, enabling the businesses to unlock more value from AI is a fun challenge. “I firmly believe the next wave of software is intelligent software,” she says. “Many businesses are only beginning to realize the real power of data and predictive technologies.”
Verta stands on that cusp, and Vartak is comfortable at the forefront of tomorrow’s possibilities.
“I want to enable that intelligent software. Even when I was at WPI, enabling technology was a theme that resonated with me,” she says. “It all goes back to innovation. There are so many new and open problems in the space and there’s a lot of room for innovation that’s going to have a pretty sizeable impact. So this is where I want to spend my time exploring.”
Tune into the entrepreneurial subculture at WPI and you’ll discover innovators with a bevy of ideas focused on solving real problems. Some aim to commercialize their ideas, others to refine a research technique, develop an academic project, or find new approaches to address social justice causes.
In his application essay, Matt Shea '03 wrote that he dreamed of sending something into space. Now 20 years later, that dream has become a reality. He joined with longtime friend Ryan McDevitt ’03 to start Benchmark Space Systems, which builds propulsion systems for satellites.