Clement Pang: AI Research, Relatable Software, and the Art of Coding

Rice CS alumnus and serial entrepreneur Clement Pang.Clement Pang (B.S. ’07) is a serial technology industry entrepreneur and investor in Silicon Valley, but when he arrived at Rice University as a freshman in 2003, the number of students majoring in Computer Science was dwindling.

“Only seven of us got a B.S. in computer science in May 2007, but the intimate nature of our small group presented unique opportunities,” said Pang. “I got my first taste of information retrieval and machine learning when Devika Subramanian invited undergrads to apply for research openings on her team.”

Pang joined Subramanian’s team as a sophomore, when the researchers were developing programs to review online news articles and trying to predict future conflicts based on that content.

“It sounded like sci-fi at the time, but we were tackling and solving really hard problems. We stayed late in the lab, then met at each other’s apartments to keep talking about the work,” he said. “To be part of a high-caliber team like that, you’d usually have to be a graduate student specializing in artificial intelligence. But Devika had three or four undergraduates and several grad students. It was a real eye-opener for me, and I worked for her
for three years.”

Both undergraduate and graduate student researchers published papers and presented posters about their work. Pang’s, published in 2006, focused on identifying duplicate and irrelevant articles. He credits his work in the Subramanian lab with a subsequent Google summer internship offer. After graduation, he joined the Google team working on fraud detection and machine learning models in real time.

At Rice, the potential applications for artificial intelligence and machine learning had captivated Pang. He completed Subramanian’s undergraduate courses and took her graduate-level courses. But it was a software design class with CS professor Stephen Wong that first revealed Pang’s start-up spirit.

“Dr. Wong’s COMP 410 was very different from my other classes. That whole semester, we worked in small teams to build tools for Dr. Dung ‘Zung’ Nguyen. Working together on a real product in a real market forced us to learn how software is actually built and to understand the philosophy of how people and projects work within teams and companies.

“We had milestones, prototypes and ship dates. That class helped us develop skillsets beyond what you usually get in college classes, and it piqued my entrepreneurial interest,” he said.

In the follow-up course, COMP 415, the class was assigned a project for an external customer, a hedge fund in another city.

“Most of us had been in COMP 410 together, but this time we worked as a large team and there was a good separation of duties and concerns by role. Some people were writing specs for the project, and we had Angela Wise working as our product manager.

“Knowing how things are built,” he said, “how a team works, and how to work faster and not just sit staring at a monitor when you get stuck – that is a unique skillset for any graduate to have. It certainly paved the way for me later.”

He enjoyed the software design course so much that he worked as Wong’s TA for three semesters. “I essentially took COMP 415 four times, and I think Angela TA’ed with me each time,” Pang said.

Although Wong has taught similar courses for more than a decade at Rice, he still remembers the projects in Pang’s class. He said, “Clement’s COMP 410 project was SkyNet, a distributed software framework that allows users to create dynamically extensible games and simulations. His COMP 415 project was a stock market simulator that HBK could use to test out their automated trading algorithms. Coincidentally, the posters for both his projects are currently hanging in my office.”

Pang loved the development process he learned in Wong’s courses, but his ultimate goal is to produce software solutions for problems that matter. “I’m a nerd and engineer at heart and I really want to create lasting, well-built software that is relatable. COMP 415 helped me realize I was looking for good problems to solve with software.

“Working at Google, founding a FinTech company and Wavefront — all those organizations are working on really hard problems. Hard math and hard architecture problems are the kind I like, not necessarily being the first to market or exploiting a niche area. I go into a problem that requires an engineering solution, then discover a market for the solution and build a company around it.”

Wavefront, Pang’s most recent success story, was a journey to create a massively scalable time series database and query engine, a SaaS product that tens of thousands of engineers rely on every day to make decisions, identify problems and find solutions to challenges they face. Having grown his company from three to more than 150 employees in six years, Pang’s job as a founder has changed significantly.

“Starting up a company, no matter how good you are at computer science or engineering, you have to be good at the vision stuff. No one tells you what to do. You have a hunch and you follow it,” said Pang.

“But as you build a large company, you are hiring more people who are centered on their individual missions, and less so on the customer directly compared to the early start-up days. Larger teams are turning out good work, but it’s less about the big picture and more about deliverables and incremental progress.

“You, as the leader, have to think of all the paths the product can take, and articulate that to the team so that they understand how it all fits and can rally behind new initiatives. They implicitly trust your judgement, so being on the management team is beyond just managing people but also how we, as founders, see the market and the need evolves, and make the right call.”

Pang feels a sense of reward and relief when his team figures out how to accomplish his vision.

“When our engineers talk to customers, read blogs, and have a sudden epiphany such that they start to come up with ideas about how to build the feature or evolve the product — that’s a good day. When they build a prototype or show me something they have looked at or mocked up, that makes me feel the team is evolving organically and the product will naturally improve.

“All of us have a little entrepreneurial spirit inside us. We all know a little about the customer, market and product. So as a tech leader and business head, I’m very happy when people take the initiative to evolve and take the product to the next level without specific instructions from me.”

He is passionate about how people code. Pang said it is too easy for engineers to feel like their programming is simply work done for someone else. “At the end of the day, coding is an art. Hence it is possible to perfect your craft. If no one was checking your work, would you still add comments, perform all the tests, make it look pretty?”

His advice to current and prospective CS students is to take ownership of their craft and their code, and allow their careers to follow: “Cultivate that feeling of excellence, want to be doing the best you can. Not to get ahead of market and not just because you get a request in your inbox, but because you want the best design for your product, and to do a damn good job at it. Talk with customers, then sit down and write good code.”