Alumni Profiles Series: Vick Khera
Vivek “Vick” Khera is CTO at Carfeine, a fast-growing tech start-up democratizing digital marketing through AI products that connect automotive retailers and buyers. He completed his Ph.D. in Computer Science at Duke in 1994 after receiving his M.S. in 1992. Since then, Dr. Khera co-founded and ran two successful businesses: GovCon, a portal for government contractors that was acquired in 1999, and MailerMailer, an award-winning email marketing tool for businesses, acquired by j2 Global (Nasdaq: JCOM) in 2017.
When did your fascination with computers begin, and how did you turn that into a career?
Oh, I’ve been into computers forever. When I started with them, they were a brand-new thing. My first computer was an Apple II+ in high school. I was always tinkering with computers and was curious to learn more about what you could do with them. I started a company right out of Duke in 1994 with my brother, focused on the internet and helping people realize what they could do on there. We then started building promotional games for companies to use as marketing. One of them involved a scratching a card—much like a lottery ticket—that you would scratch with your mouse online. That was a big success. Internet Explorer 5 was marketed with a Download and Win campaign that used this technology, and the NFL hired us for something similar. Later we ended up building a portal for government contractors, which eventually got acquired.
At some point I was sitting there thinking, well, what do I do next? I started by thinking about the problems I had running this last company. One of my big problems was that our email newsletter took me a lot of time to create, send, and then deal with all of the bounces and people asking to get off of it. My thinking was that everyone's got this problem. So, I built a whole system to manage email newsletters. We called it MailerMailer and ran that until 2017.
Did you always know that you wanted to start your own company?
No! At the end of 1994 I was home for about six months before my defense because of funding reasons. You know, I had just enough funding to pay tuition and not enough to pay rent. I found a local ISP at the time in Maryland so that I could “rlogin” [a pre-cursor to ssh] to the computers at Duke where all my data was. I would then work remotely through a terminal. I guess I was actually one of the first teleworkers, running on a 28.8 modem!
My brother is the one who has always had this entrepreneurial bug. He looked at what I was doing with the internet, and it fascinated him. He's an electrical engineer. So, he had some ideas, but he didn't understand quite how far you could take it. He convinced me to start in the business with him. It took a while. We flopped around trying to figure out exactly what to do; we didn't really have a business plan. It started out more consulting-like and then kind of stumbled its way into these two big projects [GovCon and MailerMailer] that worked out really well for us. There's no way that would happen today! We were lucky there were so few people working in the space at the time.
Since then you’ve transitioned into a role as CTO. What is that like?
Well, everybody defines their CTO differently. At some companies, you’ll be looking at very summarized reports and act as the person who collects all the information and distills it for the CEO and the board. At other companies, you will be actively working on products and diving in with the engineering team. I'm more of that level, and I think that happens more in smaller to midsize companies. That's really the kind of role I like. The fancy title is good for me, but at larger companies, it might be more of a VP or Director role.
Do you interact directly with your company’s codebase?
I do look at the code from time to time when I'm trying to figure out why we're getting some particularly curious answer, just to see if I can figure something out, but I don't create very much of it. If I do, it's proof of concept of how we should do this new algorithm or something along those lines. The AI code? Not a chance. The people on our team who work with machine learning—that is literally all they do. For the most part, that's all they can do because it's so different from the other code we create.
What skills did you learn during your Ph.D. that you use most now?
Running large numbers of experiments and collecting, sorting, organizing, and managing that data in a systemic way. That was the biggest skill that I use all the time, along with automating everything because I didn't want to hand-run 400 experiments. Back then we had a big smattering of Sun workstations around the offices and four big computers in the basement server room, plus a couple of experimental machines. I tried to spread my simulation workload as much as possible across those stations. That forced me to automate and track all the tasks. That kind of skill is very useful.
There are plenty of intangible things too. The breadth of knowledge from being around smart people and constantly learning things, diving deeply into topics. That is much harder to quantify but equally important.
Do you have a favorite memory of your time at Duke?
For basketball tickets for grad students, you have to get the season pass, right? Well, ours was literally a book of tickets. It was great. We had a single campout for grad student tickets. My friend Eric Anderson was a slightly older grad student. He was on his third career after having been an investment banker, getting his law degree, working as a lawyer for a while in New York City, and then finally deciding to get his Ph.D. in computer science. He was a Renaissance man, good at literally everything. This particular year we were camping out together.
He had his tent; I didn't have anything. He had all this equipment because he was a real adult. So, when everyone else was ordering pizza or subs for dinner, Eric cooked a three-course gourmet meal on his camp stove! People were floored! So, we're sitting there eating this beautiful, wonderful meal that I couldn't cook at home if I tried. It was just so funny and so outrageous. I think that's one of my favorite, favorite remembrances.
Ph.D. student, Computer Science
Rich Stureborg is a Ph.D. student in the Department of Computer Science. His research is focused on how to leverage Natural Language Processing (NLP) to combat vaccine misinformation. An active advocate in GPSG, Rich recently served as a student representative on the Dean-led Ph.D. Stipend Task Force. This committee ultimately made the recommendation to increase stipend rates for Ph.D. students at Duke University.