Faces of the Foundation: Michael Li
“Today’s world is drowning in data,” says Hertz Fellow Michael Li – and the glut is only getting larger. The world generated more than 16 zettabytes (that’s 16 trillion terabytes) of data in 2016, and that number will increase tenfold by 2025, according to marketing research group IDC. Li says the rapid growth of information resources has produced a qualitative change in how companies handle data – no longer squeezing as much from limited data as possible, analysts are now faced with the problem of sorting, processing, and analyzing data as rapidly as possible, often deciding in real time what information to keep or let lose.
Li is helping employers deal with this paradigm shift by providing an equivalent change in how data scientists are hired and trained. In 2014, he founded The Data Incubator, a company dedicated to training recent PhD scientists to be data scientists in the modern workplace.
Li brings a unique mix of academic and industry experience to The Data Incubator, positioning him well to take the unique path he has. As a Marshall scholar at Cambridge and then as a Hertz Fellow at Princeton, Li studied applied mathematics, solving problems in quantitative trading for his dissertation. “I’ve always been interested in the intersection between mathematics, computing, and the human part of the real world,” he says.
This humanist interest continued to show when he began working as a data scientist after receiving his PhD in 2012. The process of hiring data scientists is plagued by uncertainty on both sides, opening the door to prejudiced hiring decisions and failed hiring screens. “As a PhD student and then as a hiring manager, I saw the pain points on both sides of the interviewing table.” Li realized that he could combine his passion for teaching and his experience both hiring and being hired to improve this process.
Breaking the mold both of traditional educational pathways and of hiring/training programs, The Data Incubator’s eight-week training programs are free to fellows – they are paid for, instead, by employers who hire the trained fellows. “We’re really trying to turn the educational model on its head,” Li told Venture Beat shortly after founding the company. “The idea that you should pay for your own training when you’re so close to being employable — I just don’t think it’s right.”
The advantage of teaching data science to PhD scientists is that they’re quick learners with many of the technical and analytical skills to be highly effective data scientists. The fast pace of the program – much shorter than the master’s degree that fellows might otherwise pursue – is an advantage in teaching the most important change of mindset that PhDs-turned-data-scientists need: nimbleness. “PhD students are used to getting things done on a timeline of five years. We need to get that down to five days, if not five hours,” says Li.
In the Hertz Community, surrounded at retreats and workshops by creative people from all disciplines, Li felt inspired and enabled to try out The Data Incubator’s novel approach to education and recruitment. “In any institution or peer group there are a lot of default pathways that are subtly suggested,” he says. “One of the nice things about the Hertz is you see a large variety of very creative, very smart, very passionate colleagues engaging in a whole bunch of exciting ventures. That collection of what I like to call ‘socially productive renegades’ is a really inspiring group to be with and really helps you think about how you can achieve the most of your potential without having to conform to one of these default pathways.”