“I was always fascinated with physics and science,” says Peter Skomoroch, principal data scientist at LinkedIn. “When I was 6 or so, my parents asked me what I wanted to be, and I said theoretical physicist. I'd heard it somewhere and it sounded cool.”

During his freshman year of high school, his interest in science and data solidified when he entered a mandatory science fair. He ended up entering more science fairs and competitions in subsequent years—past the point of mandatory participation—and started leaving high school early to do neuroscience research at the local university lab.

He was studying the visual system of lizards, and how a lizard’s third eye encodes signals and information into an image. That requires a ton of data, and Peter was hooked.

“It’s kind of like cracking a code, coming up with a model that actually replicates something real,” he says.

He went on to study neurology and biology at Brandeis University, but realized he didn’t like lab work and was more interested in analyzing data, so he ended up graduating with a degree in math and physics.

After he graduated, he went on to work at a startup called ProfitLogic, which takes large amounts of transactional data from retailers and models the life cycles of actual fashion goods.

“Working with large data sets from Walmart or JCPenney, you really get to cut your teeth on the grittiness of real world data,” he says. “You can actually model the growth curve and decay of jeans or shirts or handbags, and you can decide how much to charge for those products if you understand the elasticity and the nature of the product."

In 2004, after a few years at the startup, he returned to the classroom to brush up on his machine learning skills with some graduate coursework at the Massachusetts Institute of Technology “I wanted to learn more about those algorithms and how to operate at scale,” he says.

“If undergraduate education is successful, it teaches you to know what you don’t know,” he says. “A gap I saw was, there’s this field of machine learning and pattern recognition developing. That was a domain area I wanted to get more expertise in.” He’d encountered problems of having millions of curves and needing to classify, group, and treat those curves. “When you’re looking at those curves manually, you know there’s something wrong and you turn to algorithms to solve it.”

Though he bailed before finishing a Ph.D., he spent his year at MIT studying neural networks, machine learning, and math. He did research in the biodefense group at the MIT Lincoln Laboratory, working with an interdisciplinary group of scientists working on low-cost sensors that can detect airborne pathogens.

After that, he moved down to D.C. because his wife was living there, and got a job as a senior research engineer on the AOL search analytics team. “That was my first chance to really work with Internet data, which has its own unique intricacies,” he says. “It’s almost getting at psychology, in a way, the idea of understanding who someone is, what they’re looking for, and how you can help them.”

As he’s moved out to Silicon Valley and joined LinkedIn, that’s something he’s continued to be fascinated with. “Given a small amount of information, can we help you understand better what are your strengths, who are you like, if you want to be like someone else how do you get there, and how do you get the job you want?” he says. “It’s the more human side of being fascinated with the mind and how it works: Who are we, and how can data tell us who we are?”

During his time at AOL, he founded a website called Data Wrangling, which started as his blog but now offers software development services for clients needing scalable data mining or search applications. The site helped him to build up a network in Silicon Valley before he even moved to the area, which helped him land the job at LinkedIn.

“There was this huge amount of data and there was so much potential, that’s what drew me to working at LinkedIn,” he says. “It was 2008 [when I started], and the economy had just taken a big hit, and I couldn’t think of a better place to actually help.”

His biggest project at LinkedIn so far has been the Skills product. “When I came here, the data group was just getting going,” he recalls. “One of the things I saw at the time was we have all this great information on companies, job titles, education and schools, and one aspect that was missing for me was skills. It might seem analytical to think of people this way, but when I think of identity I think, what is the vector of skills, interests, and topics that represents you?”