Is a PhD absolutely required to become a researcher in industry? Many people would say yes. However, when we asked this question to YuLing Chen, who’s a production engineer and data scientist on the Meta Network Infrastructure team, her answer was “not necessarily.”

Over the course of her career, Chen has published academic papers and was the inventor on a series of U.S. patents, earning what many would consider a “PhD equivalent.” Chen believes that while earning a PhD is a tremendous accomplishment, having one doesn’t need to be a hard prerequisite to conduct industry research; in fact, non-PhDs could actually help contribute to the diversity of thought in the research community.

“Diversity and inclusion needs to be extended not only to traditionally underrepresented people, but also to non-PhD students and professionals in research fields,” says Chen. “High-quality research that generates impact to products and services requires curiosity, creativity, continuous learning, problem solving, hands-on prototyping, and lots of persistence. Anyone with these capabilities can be successful in applied research in industry.”

We reached out to Chen to learn more about what it’s like to be a non-PhD conducting applied research in industry, her recent research projects at Meta, and how she overcame challenges both as a woman in the tech field and as a researcher without a PhD degree. She also offers advice for non-PhD students and professionals who are interested in applied research in industry.

Q: How did you first become interested in research?

YuLing Chen: Both of my parents were researchers in geology in China. When I was little, I was encouraged to explore the world with curiosity. During my study of computer science at Nanjing University, I liked to ask questions so much that my classmates called me “the girl with 100 thousand why questions.”

Innovation and research play important roles in the high tech industry. While working at several companies during my career, I have been persistent in exploring new technologies to create solutions to both existing and new problem domains. In 2014, I proposed a highly scalable distributed data collection, processing, and storage system to handle large-scale networking telemetry data, which became the first U.S. patent in my career.

Q: How did you end up where you are today? What did the path look like for you?

YL: I have had a passion for exploring new areas and subjects since I was young. Despite my broad range of interests in biology, chemistry, literature, and arts, I chose computer science as my major because it was brand new back then, and my dream school, Nanjing University, excelled in computer science.

I’m also an experimentalist, and I always enjoy seeing the results generated from the products that I build. Therefore, instead of pursuing an advanced degree in college, I went to work in the computer industry right after my undergraduate studies. Seven years later, I felt that I needed to catch up with the latest technology, so I went back to school and earned a Master’s degree in Computer Science from McGill University. After graduation, I worked for a series of companies in different domains including IBM, Ericsson, Dell, and Cisco. Three years ago, I joined Meta working as a production engineer with a focus on leveraging large-scale network data to improve network reliability, performance, and efficiency.

Q: What challenges have you encountered doing research without a PhD? How did you overcome them?

YL: As a woman without a PhD, I am highly underrepresented in computer science-related research fields. Back in my time as an undergraduate, I remember there were only two women in a class of 30 students. Even in my current day-to-day work, most of the time, I am the only woman in the meeting. It makes me feel out of place, and the fear of being misunderstood makes me doubt myself when presenting technically in front of people who are all of a different background. In one of my applications to a research position working under the Chief Technology Officer of a renowned network company, I was clearly told that without a PhD, my application was not going to be considered.

Despite all these challenges, I knew I wanted to participate in and contribute to the knowledge community, so I persistently pursued opportunities in research and innovation. I kept up with the latest technologies by attending conferences, reading up on research papers, and applying new technologies in my daily work. As of now, I am the inventor on nine U.S. patents and published three papers in different areas including network service orchestration, network function virtualization, software defined network, blockchain and distributed ledgers, and, recently, machine learning for creativity and design.

Q: What are some of your most recent projects and research at Meta?

YL: After I joined Meta, I focused on developing data-driven solutions to help with CDN reliability, performance, and efficiency. This includes predicting peak traffic volume for accurate NYE capacity planning, continuous data center capacity headroom modeling and monitoring for reliable delivery of Meta user traffic, as well as user performance optimization using data mining and machine learning algorithms.

I am also very passionate about new and emerging use cases such as AR/VR being deployed on the Meta network infrastructure. In order to understand the AR/VR workload where Generative Adversarial Network (GAN) models are widely used, I collaborated with UC Berkeley and developed a GAN-based application that can generate impressionist landscape paintings based on user input. In December 2021, our paper “GANArtworks in the mood”was accepted and published in the 5th NeurIPS Machine Learning for Creativity and Design.

Q: What advice would you give to female or non-PhDs trying to get into applied research fields in industry?

YL: My first piece of advice is to keep up with the latest technology and apply it to your day-to-day work. Read research articles, go to technical conferences, take courses online to expand your knowledge and technical skills — you might even end up wanting to go back to school! Three years ago, in order to learn machine learning algorithms in a systematic way, I started my second master’s degree in data science at UC Berkeley. This experience turned out to be very rewarding and helped me to submit two papers in machine learning and data science areas.

Second, try to work on projects that afford you more opportunities to use new technologies. At Meta, especially as we’re working on building the metaverse, we will see more projects from both application and infrastructure levels that would require us to apply new technology with creativity. Get involved in these projects, because there are plenty of opportunities for research and innovation.

Finally, be persistent. When you explore research areas with new ideas, it’s easy to doubt yourself sometimes. As a non-PhD professional, you will encounter more challenges than others, especially if you’re an underrepresented person. Be bold and trust yourself. Show people the results with data that support your ideas. Treat challenges and doubts as the normal process of pursuing innovations. A good idea will always be recognized sooner or later.