As a Rice University freshman focused on bioengineering, Drew Bryant asked Lydia Kavraki about summer research opportunities. Kavraki is the Noah Harding Professor of Computer Science, but is also a Rice professor of bioengineering, mechanical engineering, and electrical and computer engineering.
Bryant spent several years working in her Physical and Biological Computing group, but Kavraki is also well-known for her Robotics Lab. Ironically, Bryant’s first role after graduation involved programming a laboratory robot.
“After graduating, I opted for a six-month internship before deciding if and when I wanted to go to grad school,” said Bryant. “The protein purification lab that hired me happened to be testing robotics replication of the scientists’ experiments. There was this brand new robotics system that could pick up trays and move them around — a very nice machine, but no one there knew how to program it.
“I said, ‘I can do that’ and even though I was only an intern, I was able to program the robot to replicate the experiments and also cut the time down to six hours instead of the two or three months it took the scientists to perform by hand.”
Bryant said his experience with Kavraki had given him confidence in treading both worlds – bioengineering and computer science – but it was the lab robot programming that settled his direction for graduate school.
“I was involved in one of the California lab’s first big robotics tests, replicating the experiments and speeding up the results. Everyone was excited about the outcome and that the results were reproducible. That’s when I thought, ‘there is a whole career in this!’ and realized I was actually enjoying the robotics part more than the work in the lab.”
He felt drawn to data analysis and computational tools, so he contacted Kavraki about returning to Rice as one of her graduate students. He said, “Lydia was happy to have me back working on 3D protein structure analysis. We were trying to predict the function of unknown proteins.”
Bryant enjoyed returning to that familiar research theme, which had already provided material for several publications and an award for creativity in research. His return to deep research in his former team also allowed him to complete his CS master’s and Ph.D. requirements in just four years.
Unfortunately, those four years began just as effects of the global financial crisis started rippling through almost every industry.
Bryant said, “When I was graduating with a Ph.D. and a mix of BioE and CS skills, my hope had been to go back to a lab in a bioinformatics role. I kept applying for the limited roles in that area while watching the tech industry heat up. The tech opportunities greatly exceeded opportunities in bioinformatics, so I began browsing through those job postings.
“An Amazon job listed bioinformatics as a preferred background skillset. As we began discussing the role, I came to realize that a lot of pattern recognition is based on informatics. The interview went well and I got an offer that took me to Seattle.”
His job at Amazon introduced him to the industrial version of machine learning and data analysis, and he thoroughly enjoyed the cloud-based environment. After hitting his stride with the tools and systems in Amazon’s arsenal, he began wondering if he could create his own.
“I realized there was still a lot of opportunity to refine machine learning and data analysis at cloud scale and began researching companies that specialized in that field, like Facebook, Google, and Tableau.
The Google offer included the opportunity to start green field, fresh. They were adding to their platform and I would be able to use tools like those I’d worked on at Amazon, or build my own and build them specifically to work in Google’s cloud environment.”
He spent several years working on the cloud data team, including collaboration with Google’s cloud genomics team. Then Google launched a new group focused on climate and energy research. Bryant said his colleagues in the genomics team remembered his background in protein research and passed his name along to the new team.
“They were going to start by looking at protein research and my background was a strong match. I saw the opportunity as more of a research role, where I could apply machine learning at scale on proteins in silico, using the domain knowledge I had acquired at Rice.”
Bryant continued working in Google’s Seattle campus even though most of his new team would be based in Mountain View, California, along with their collaborators in the Google Accelerated Science team.
The remote partnership with his colleagues doesn’t bother Bryant, who arrives early to exercise at the company gym and eat breakfast on campus. After diving into coding for several hours, he takes a break and reads through one of the research papers stacked on his desk during lunch.
“Then I’ll dial in for our daily video conference with colleagues in Mountain View late in the afternoon,” he said, “but we’re also in touch through our internal chat system. Then there’s more coding in the afternoon.”
Bryant attributes much of his success to his Ph.D. training at Rice. “Getting a Ph.D. is relevant to industry roles because it prepares you to navigate completely on your own,” he said. “You learn to figure out a plan when there isn’t any plan, and to find the frontier and then ask questions about it.
“You also understand how to structure your projects so you show progress rather than appearing lost. In addition to not being afraid of being out in the wilderness, you understand how to take current knowledge, extrapolate from it and determine how to know if you are on right track.”
His favorite part of the new role is talking with external collaborators at various universities. Bryant said, “We get data from their labs to run, acting like machine learning consultants for labs that have a lot of data but not a lot of scale. How can we use our large scale infrastructure to help answer questions about your data? When we can help them get answers about their data they can’t get on their own, that’s a great day.”