Greg Kuhlmann was one of two data scientists that Apple brought in to find new approaches to preventing fraud. That was back in 2011, and Kuhlmann was working at Apple's North Austin office, which has since ballooned into a massive campus.
Before he came aboard, he said the main focus was gift certificate fraud. But with its rapidly growing App Store, it needed more aggressive and efficient ways to identify malicious behavior. Soon, games began offering tokens and ways for users to earn cash.
"Suddenly there's a path to cash for fraudsters to basically take stolen credit cards, buy those things, resell them at a discount, and suddenly they're just like making tons and tons of money off of Apple," Kuhlmann said. "So the fraud problem exploded, and they were astute enough to recognize that machine learning is a way that they can attack this problem."
Kuhlmann, who left Apple in 2019, co-founded a startup with Lucas McGrew in 2020 called Sumatra that aims to give more businesses access to advanced AI and ML methods to not only fight fraud, but also developing personalized shopping experiences based on real time user activity. McGrew was most recently a senior software engineer at DRW.
The startup has now raised $1.5 million in seed funding.
Sumatra did much of its fundraising before the venture market began to take a more cautious approach amid falling valuations and huge decrease in tech IPOs. The co-founders sought out Austin venture capital firms because they knew local VCs would understand the problems they're trying to solve, as well as one of their earliest customers, Austin-based insurtech startup The Zebra.
That led them to meet with leaders at Austin-based firm S3 Ventures.
"We actually talked a couple of times over over beers, and that was enough for us to get commitment on on both sides," Kuhlmann said. "And it's funny – afterward we actually kind of closed the deal, we talked about how for them this was the first time they'd ever actually closed a deal without seeing a pitch deck. It was just like through these conversations that it worked out."
That doesn't mean they didn't use their pitch deck. They presented it to other S3 Ventures team members, as well as other firms. Eventually, their round came to include Sentiero Ventures, Data Power Ventures and AI Sprouts.
Sumatra's platform helps data scientists at all levels more quickly deploy machine learning and analytics that help with fraud, conversion optimization and delivery logistics.
Kuhlmann attributes Sumatra's early success, in-part, to his former boss at Apple, who introduced him to one of his ultimate frisbee buddies, Meetesh Karia, who is chief technology officer at The Zebra. That set the stage for Sumatra to begin working on customer experience solutions with The Zebra. While Sumatra is focused on detecting fraud, its AI and ML capabilities also allow it to automatically monitor and react to customer behavior and personalize shopping experiences.
That led to Sumatra's first big use case.
“The platform has enabled our small data science team to ship multiple real-time ML and analytics services. Additionally, with Sumatra, our team is quickly able to leverage the tool for new use cases across the organization,” Karia stated in a news release.
Sumatra also works with Lemon Squeezy, a digital products platform, although it declined to name other customers due to confidentiality agreements.
While Kuhlmann said Sumatra raised money before VCs began to make deals more cautiously, he said he's still girding for tougher times.
"It's definitely the case that it it makes us make sure that we are controlling our burn and recognizing that the next round may not come nine months from now or even 12 months from now," he said. "We need to plan for something closer to 18-plus months."
The startup is poised to bring on its first full-time employee, and it has open roles for engineers. It plans to hire for its sales team in the coming months.
"We believe that people want to do really interesting work, and the message that we've been putting out there has been to kind of talk less about the kombucha on tap and the Ping Pong tables and things like that and focus more on the work itself," he said. "The area that we're in has the combination of every cool computer science problem that anybody would ever want to solve."
