Steven Grimm

Ikasketak

B.A. computer science, UC-Santa Cruz

Prior jobs

Sun Microsystems, Enterprise Integration Technologies, HearMe, etc.

You've been at a lot of Silicon Valley companies. What brought you to Facebook?

I had worked with Facebook vice president Jeff Rothschild at a different company in the mid-1990s, so when he called me in 2005 and asked me to do some contract work for Facebook, I was intrigued. Once I started, two things really impressed me. First, the brainpower of Facebook employees was very high. It was one of the highest concentrations of smart people that I'd ever seen. Second, the site was growing so fast, and touching so many people, that anything good you did would reach a wide audience just about instantly.

What have you tackled at Facebook?

I came in with low-level networking knowledge, so I started by doing a lot of work on improving speed with database performance, via memcached. I spent about 1.5 years on infrastructure projects.

Then I got interested in translating the site. I was studying Chinese at the time, partly because the contrasts in different languages' grammars is a good brain workout. Something as basic as saying that two people are married is done quite differently in various languages. We wanted to translate Facebook into dozens of languages, which would involve lots of different challenges. I like being able to combine a big-picture perspective — where you're thinking about how systems fit together — with a chance to pursue specific challenges in great detail.

How much freedom do you get in your day-to-day work?

Direct management here can be invisible, and not in a bad way. You always need to manage the people who need a lot of supervision. But there also are engineers that can be left alone, and that's who we try hard to hire. It's more of a research-scientist model. The managers' role is to make sure people can work effectively on the right projects without being impeded by anything. Facebook lets that happen.

We're also encouraged to work on open-source projects wherever it's appropriate. All the work I did on memcached was contributed back to the public code base, and if we're working on something brand new that's a good candidate for opening up, there's no problem taking the extra time to get it packaged and announced to the world.

What about top leadership?

Our CEO, Mark Zuckerberg, can be pretty hands on. But he picks his spots. On the internationalization project, for example, we met with him to review the user interface before we launched. He looked at it and said: "That's reasonable." We had the freedom to go ahead with what we thought was the best approach. I think his insights have proven generally right. The success of our site speaks for itself.

How good is Facebook at getting projects finished?

We're not in a situation where we've made promises to external customers, to the point that we have to go on death marches to get things done on time. It's pretty reasonable here. We're trying to deliver great products to our end users, usually without having told them ahead of time what's coming. So we can adjust completion schedules or even change the specs if it's obvious that's the right thing to do. We have more control.

Are you comfortable being the voice of experience?

I'm not the only one. Probably a quarter of the engineers have as much experience as I do. Our engineers have all kinds of backgrounds; some are fairly recent graduates while others have been working for 15 years or more. We try hard to keep the environment egalitarian; if you can make a solid technical case for something, you'll convince your peers, whether you've worked at 10 previous jobs or none.

Does Facebook plan — or improvise?

It's pretty dynamic. You never quite know a month from now what you're going to be working on. Someone gets an idea and runs with it. That's encouraged. There's no 10-year plan here, where everything has to conform to a strategy that's been locked in for a long time. If there were, half the engineers here would probably quit.