Sean Robinson, of Pioneer Square Labs, Talks Data Science, Machine Learning, and AI
Sean Robinson began his career developing algorithms for the NASA/DOE Gamma Ray Large Area Space Telescope (GLAST) after receiving his Ph.D. in computational physics from the University of Washington. A self-described “recovering academic,” Robinson spent 12 years as a nuclear physicist before his professional focus shifted towards data science. His current role encompasses research and project management, utilizing his experience with grant-, agency-, and privately-funded projects to fuel his start-up studio toward continued success.
Pioneer Square Labs, Seattle, Washington
As Pioneer Square Labs’ Principal Data Scientist, Robinson supports the data science, machine learning, and artificial intelligence (AI) needs of his company’s varied client base. Robinson is a steward and cultivator of new ideas, gauging the technical feasibility and market opportunity of potential concepts, brainstorming “tons of ideas” and reducing them to only the most promising and profitable. After assembling a core team to back the new idea, Pioneer Square Labs seeks investment capital to grow the emerging business; it’s then the new company takes the reins.
Robinson explains that in the realm of scientific research, even the best ideas require extensive, upfront legwork before funding offers materialize. Agency and grant funding often come with restrictive conditions and are not intended to sustain a project long-term. Conversely, private funding ensures flexibility in building a robust, enduring business. “I don’t necessarily have to…procure anything other than the support of immediate peers and perhaps some of their time and interest to develop an idea.” By securing local investment and community funding, as opposed to pursuing large grants, Pioneer Square Labs can begin projects immediately after conception. In Robinson’s words, “[this lets] them fail; which is an important part of finding the things that won’t fail.”
“At a place like Pioneer Square Labs, researchers, engineers, and business development people can all come together and…simply go ahead and do experiments” without the burden of uncertain funding.
The Ideal Data Scientist
Robinson compares the future of data science to the atmosphere of late-90s web development. “These were times when eCommerce was in its infancy…everything was very wild and wooly.” At the time, computer science-specific curriculums were rare, despite the demand for computer science workers, and soon-to-be big-name companies took anyone with relevant, transferrable knowledge. “Basically every nerd with a pocket protector, myself included, who was doing something else–maybe occupying a corner of a dusty lab somewhere or involved in a relatively unpopular topic in their college program–could suddenly hold down a fairly high income.”
He explains data science and AI learning is following a similar trajectory: the demand is growing, college programs and stringent standards are not yet in place, and the field is primed to welcome passionate individuals from a myriad of career backgrounds.
According to Robinson, success in the emerging data science fields begins with a solid understanding of mathematic, statistical, and analytical basics. As a fledgling field with minimal standardized curriculum, Robinson believes success begins with a well-stocked mathematical toolbox, an inherently inquisitive nature, and a passion for creative problem solving. “Do you really love puzzles? If somebody gives you a riddle, a puzzle, or a paradox, does it just light you up? If you are fascinated by the topic [of data science and machine learning] and by that kind of problem solving, then we’re in a place right now…where that’s a fundamental job skill.” Robinson further highlights Bellevue College’s specialized data analytics degree, as well as their new Robotics and AI programs, praising their willingness to explore a growing trend.
“Right now the field is very young. It is very broad. There’s a lot of room for creativity and that creativity is fueled by your desire to explore the world, the world that you’re working in.”
The Future of Data Science
Robinson is certain machine learning as a service is on-track “to be gigantic.” Facial recognition, image analysis, text translation, and natural language processing techniques are advancing in increasingly sophisticated directions while simultaneously becoming more accessible. He believes process automation will continue to refine and streamline technological processes, creating standards that eventually negate the need to constantly create custom algorithms. Robinson believes present-day tools will eventually transcend “anything that an individual would trivially develop.” Optimization of, and training around, existing systems will lay the groundwork for those data science tools just emerging on the horizon.
Robinson encourages prospective data scientists to become intimately familiar with the tools available now, as they have already been laid as the foundation for greater advancement. “Now is the time to really look at the tools that we have now…[and] really understand the way that they work, such that when the [future] tools show up, you’ll…understand what’s going on under the hood.”