When Cindy Gonzales joined LLNL as an administrator with the Computing Scholar Program in 2016, she knew the Lab was a place where she could build a career. Her husband, Mathew, works in Computing’s Information Technology Operations Division. Her father-in-law and other extended family members also work at LLNL. Previously, Gonzales was an apprenticeship coordinator in the construction industry and had taken some classes toward a bachelor’s degree in liberal studies.

Just a few months into her LLNL employment, Gonzales attended a machine learning (ML) seminar given by staff scientist David Buttler. She says, “I was a part-time student and full-time employee when I learned what ML was and how other disciplines, like the statistics course I was enrolled in, applied to it. I thought, I could do this.” She switched her major to statistics. The rest is history.

Through LLNL’s Data Science Immersion Program, Gonzales is now among the Lab’s newest data scientists. For two and a half years, she juggled a demanding workload—coordinating Computing’s Scholar Program, interning with data scientists, learning from mentors, supporting LLNL’s Data Science Institute (DSI), and attending college part time—while also having her first child. Today, she is a data scientist in LLNL’s Global Security Computing Applications Division and uses ML to detect objects in satellite imagery.

Gonzales explains, “Data science is a diverse field, which makes it both exciting and challenging. You need a background in many different areas, such as computer science and statistics, plus domain knowledge. These skills will open doors to other scientific domains.”

Education and Immersion

Gonzales attributes her successful career transition to many factors. At the forefront is LLNL’s Education Assistance Program (EAP), which has enabled her to finish a B.S. at Cal State University’s East Bay (CSUEB) campus and begin a distance-learning M.S. program at Johns Hopkins. The EAP helps employees with tuition expenses toward certificate programs and academic degrees.

Computing workforce manager Marcey Kelley states, “Our employees are our greatest resource, so it makes absolute sense to invest in their career development. We strongly encourage our employees to take advantage of this great opportunity.”

Classroom study is one thing; on-the-job experience is another. DSI director Michael Goldman teamed up with Kelley to launch the pilot Data Science Immersion Program, which granted Gonzales dedicated internships with LLNL experts in her newfound field. “The Lab is a unique place, and we often find ourselves working in areas we never envisioned,” notes Goldman. “Investing in our workforce maintains our ability to perform cutting-edge research. The Immersion Program is one way of providing the mechanisms for people to contribute to these new areas.” Buttler agrees, “Real projects at the Lab are always much more complex than any classroom setting.”

Last fall, Gonzales interned with Buttler on a text-processing project. With no programming background, she learned the fundamentals of Python, Linux, and linear algebra. Buttler explains, “Cindy rolled up her sleeves and dug in. She developed techniques for using word-embeddings for text analysis and semantic graph generation related to materials synthesis. By the end of the three-month internship, she had created a classifier that outperformed a system we had been using to classify sentences.”

In February, Gonzales began learning about computer vision under the guidance of data scientist Sam Sakla. She worked on image classification and segmentation problems, practiced training neural networks, and developed an ML algorithm to detect clouds in satellite imagery.

Sakla elaborates, “This project allowed Cindy to perform exploratory analysis on a dataset, format the data so it can be used to train an ML algorithm, train the model, and evaluate its performance both qualitatively and quantitatively. Such deep neural networks are now being used ubiquitously across the Lab for many applications.” Of the Immersion Program, he adds, “Being involved in a project, coding, and performing the experiments is completely different from reading about data science and ML in a textbook.”

The Data Science Immersion Program was so productive that Gonzales’s plans changed again. Sakla hired her full time before she could start another internship project. “We needed someone who was familiar with deep learning for computer vision applications and someone who could code and perform the experiments. Cindy happened to be a great candidate,” he says.

‘You Can Do It’

While in school, Gonzales heard the conventional advice—“Gain experience at a Silicon Valley company, then come back to the Lab” and “You need a Ph.D. to pursue a technical staff position”—and rejected it. “There has to be a way I can do this,” she recalls thinking. She networked at DSI events and spoke to data science colleagues like Buttler, who points out, “There’s no single career roadmap for an employee to follow.”

Adjusting her mentality was another challenge. “I was good at my previous career,” Gonzales notes. “It can be hard to adjust how you think of yourself in a new area. You might be scared to fail, but the first step is to reach out to people who will help you.”

Gonzales credits her family, professors, supervisors, and mentors with supporting her education and career. She advises, “Find your champions. Take advantage of the EAP’s resources. Even if you are not currently working in the field you want to be in, even if you can only take one class per semester, just be persistent. You can do it.”

Within a few months of restarting her undergraduate studies, Gonzales was awarded CSUEB’s Justin Randle memorial scholarship for statistics students. She also received a scholarship from the Lawrence Livermore Laboratory Women’s Association. This fall will mark another impressive accomplishment: Gonzales will share her and Sakla’s work at the Applied Imagery and Pattern Recognition workshop—her first technical publication and presentation. These experiences have inspired her to pursue a master’s degree in data science.

Goldman asserts, “Cindy joined an intensive project, then learned about the domain, tools, and techniques, then performed the research and achieved results, and then wrote a paper that was accepted at a major workshop. These are phenomenal achievements that demonstrate her tenacity and drive.” Kelley adds, “Cindy’s positive attitude, hard work, and dedication have made her successful.”

As she settles into her new position and graduate degree program while raising a toddler, Gonzales acknowledges the difficulty of balancing career development, graduate school, and family. The situation motivates her to finish her M.S. before her son enters kindergarten. She says, “Transitioning from one job to a completely different one is uniquely challenging. Take it one step at a time. Looking back to the start of this journey, I’m happy things lined up the way they did and that I had the mettle to continue."