Ruth Huang Miller has woven academic and industry experience into a technology career spanning roles like computer science professor, medical school researcher, and entrepreneur.
Currently, the Rice University Computer Science alumna (MCS ‘99) is a data scientist for Panera Bread and she credits Rice’s graduate program for setting her up for success across a variety of roles. Rice is where she began expanding her capacity to acquire knowledge.
Miller said, “I remember Moshe Vardi taking all the grad students to lunch at the Faculty Club and talking to us about what we learned. It really stuck with me –Rice taught us HOW to learn. Knowing how to learn is much more important than any of the specifics taught in a class.
“As a research scientist, I focus on data science and database work, but I also spend a lot of time in compilers and algorithms. To look at algorithms means to also take other factors –like complexity and timing– into consideration. That is something I learned at Rice that has really helped my career.”
She said the ability to consider many different components of a problem applies to almost every challenge she’s encountered, including teaching introductory programming to college students.
“Beyond recognizing a good algorithm, I also consider indications that might indicate a bad habit or when someone is going down the wrong path. Each semester, I reminded students that we do actually look through their code to identify both good and bad craft.”
Miller said the current big data push is also dependent on knowing how to write algorithms, a point proved by her previous roles in companies that were charged by their quantity of CPU and storage usage.
“When we have these multi-core cloud computing environments, we really need core algorithms to compute smartly. I can write an inefficient data query that takes 10X more resources than an efficient query. Yes, it is going to directly impact the number of seconds we are charged. But worse, in shared environments, an inefficient algorithm can actually break down our whole cluster.”
She worked in industry following her Rice degree, then leveraged her Ph.D. in CS from the University of Houston to pursue an academic career. After teaching and researching in private universities for 15 years, Miller switched back to industry.
“Faculty members should always be aware of what’s going on in the tech world,” she said, “but in my case, I began to feel left behind. My students would share the kinds of questions their employers were asking or tell me where they had gotten offers, the kinds of project they would be working on, and I had a growing itch to be solving those problems myself.
“When some of our alumni brought projects to work on in alliances with our school, I actually became envious of their projects! I knew I either had to start my own company or go back into industry to work with those big data sets again.”
While retaining her faculty position at the University of Washington in St. Louis, Miller co-founded a startup and the partners began seeking investors to turn their research into a funded venture.
Miller said, “We participated in a couple of entrepreneur workshops and grant challenges, and we met with investors –but the cost of those early investments was too high. Early money is the most expensive, and we needed more deliverables to get investments at a lower cost to our future returns.
“I’m still part of that startup, but we had to take some of the activities off the table and our process stuttered when one of the partners began supporting a family member through a health issue. St Louis has a vibrant startup community, but I chose to work full time in another company and gave up teaching.”
She looked for an established environment where she could contribute immediately but also re-learn design and decision-making processes. Miller identified and accepted the tradeoffs, like joining an organization with strong learning and growth potential for its employees, but with more rigid work hours than she’d enjoyed in academia – or sometimes working overtime instead of spending her leisure time working on the startup. At Panera Bread, she found a wealth of data to examine, and she began to thrive again.
The roles of data scientists and data analysts may have evolved since Miller completed her PhD thesis in data mining, but graduate school had showed her how to commit to a problem without a known solution.
“A data scientist writes queries and performs summarizations, then goes beyond the obvious and looks for insights. If they find new insights, they also have to determine if those are relevant. Although it is titled ‘scientist,’ the role has fewer formal methods and steps to follow than our colleagues in the natural sciences. Data scientists need to have more peripheral awareness and a knack of imagining a path beyond the straight line of from here to there,” said Miller.
She said data analysts often take existing knowledge and procedures and re-run numbers, altering queries to get slightly different results or reinforce earlier results. Conversely, data scientists focus on the unknown.
“Data scientists have to figure out something they don’t already know. Getting a Ph.D. helped me develop stamina for the long haul. You don’t have to get a Ph.D. to learn how to look outside the box and tackle problems that haven’t been solved before. But somewhere, some time, you have to spend long hours – and days and weeks – working on problems no one else had developed a method or formula to solve. Pursuing a Ph.D. gives you space and time to take that deep dive.”
But she also values the professional skills she developed in the Rice MCS program, saying the two types of graduate programs address separate goals.
“It’s like having two boxes of skills and experiences,” she said. “You can build a box and fill it up with MCS courses – great courses introducing valuable skills, things you can take into your profession and spend more time there improving and increasing those skills as you go.
“But there isn’t enough time in the professional master’s program to spend weeks ‘not making progress’ on a research problem. Going for a Ph.D. allows you to spend more time learning how to build the box or how the boxes have been built and how you might build them differently.
“You have the chance to take them apart and put back together in a different fashion. That luxury of time and space to take apart the boxes and reassemble or perhaps even merge them is what gives you the confidence to go into future unknown problems with the expectation that you WILL find a solution that hasn’t been revealed before.”
Miller encourages current and prospective students to continue learning, whether it is in a graduate program or in their career.
“A good day at work is when I am learning something new,” she said. “The day I stop learning is the day I will be falling behind in my profession.”