IBM's new CEO Virginia Rometty attributed much of her rise to taking on tasks she wasn't ready for, in essence growing into the job. Yet companies increasingly don't allow that. They want job candidates to have the exact skills needed when they walk in the door - and that only exacerbates the difficulty in finding the right talent, wrote Peter Cappelli, a professor at the University of Pennsylvania's Wharton School, in a recent Wall Street Journal article.
An exception is the Pittsburgh startup Pikimal where CEO Eric Silver has been experimenting with creative ways to find developers. Pikimal pools article reviews, images and other data to help people shop smarter, a process that involves writing code to generate machine-written text. Pittsburgh Business Times ran a piece about Michael Hoffman, a graduate student in analytical philosophy and mathematical logic, who was looking to change directions. Pikimal hired him, where he became a Ruby developer.
I tracked down Silver to learn more about this and found that the company of 18 employees has since taken on two other non-traditional developers: a second analytical philosophy student and one who studied political science at Stanford. He told me Hoffman initially took an internship in content development, a position he was overqualified for, worked extremely hard and taught himself to code. Explained Silver:
We started understanding that a lot of what he'd been studying as a philosophy student was really close to the semantic search system that we've been trying to build. ... We do a lot of work to structure information and data, so what for us is a content-production position ends up close to what you might expect from programming from a semanticist. It's a different set of skills from writing an article. It's going through an ontology of data and identifying what the hierarchical relationship is. In some ways, it was an ongoing interview to determine whether he had this other set of skills. And he really stepped up with a passionate intelligence that's rare and admirable.
Hoffman started full time at the company in May. Silver said that hiring Hoffman didn't seem that big a risk, probably due to his own experience. Formerly chief marketing officer for startup ModCloth, Silver decided to start his own company. He realized he needed a prototype to show potential investors and learned enough coding to get a site up, attract investors and a couple of programmers.
So I've seen that path that a person can, with a certain amount of freedom and short-term goals, figure out what they need to do to make that work and can eventually learn to do that well. So it wasn't as much a risk for me as for someone who hadn't had that experience.
At one point, the macro software that she'd written would open up Excel and use Excel to do the functions that she needed, then close Excel. Then one day she realized that opening Excel was really slowing down the program, so she recoded the functions that were being used in Excel into the macro software. And that was a moment when [we realized that] clearly this is a person who can become a programmer. So we got her into the code base, gave her a set of programming tools and moved her into that role as well.
With Carnegie Mellon and other top schools nearby, it's not that the area lacks IT talent, Silver told me, but those graduates are hotly recruited by Silicon Valley companies paying top dollar. And you have to start recruiting about a year before you really plan to make the hire. Neither is really practical for a startup, he said.
He described his staff this way:
In the company, there are two sorts of individuals. One is those people who have followed a traditional route, 3.8 GPA or higher from excellent schools, and the other we have is sort of a brilliant hacker, somebody who has often dropped out of programs, not because they couldn't keep their life together, but because they couldn't understand why what they were learning was important or how it was relevant to the world. Both of those folks have been very successful here. I've found it extremely helpful for a startup to have people around who have a hard time doing what they're told if what they're told to do is the wrong thing.
There are situations where it would not be feasible to let workers learn programming on the job - a prefunded startup, a large company that needs to show immediate results, "but I think we landed in the sweet spot," Silver said. And he conceded he might not be so open to learning on the job in less hard-to-fill positions, such as public relations or marketing. As to what the rookie programmers are paid, he said:
We overpay them for a period of months and then we underpay them for a period of months. But once they have their skills, market forces are what they are. There's no deception here, there's no slavery. In the end, their skills are their own. ...
... you can think there will be two to four months where they will not be as productive It's comparable to a headhunter's fee, the cost of a recruiting program to find that person. I think when you're able to do it well, people are thankful for the opportunity and security to pursue a new set of skills. It's really worked out well for us. It's definitely a gamble we would make again.