Meet the data scientist plotting Twitter’s revival

240 hopefuls gathered in a room in Singapore and listened to Twitter executives give a spiel about the company. Within the crowd were data scientists and PhD candidates with a combined IQ far surpassing the number of users on Twitter.

Okay, I’m exaggerating.

But this is true: the social network needs all the smarts it can find. Twitter is trying to get its mojo back. It’s been stuck at just over 300 million monthly active users for around a year. Not a good look for a Silicon Valley company that wants to catch up with Facebook and fend off Snapchat.

Linus Lee, however, is unperturbed. And judging by the crowd that gathered, he may be onto something.

The senior data scientist at Twitter is returning home to build a data science team right here. He’s the Singapore son who’s answered his nation’s clarion call to bolster its tech ambitions.

The pool of 240 will be whittled down to 10 – a task force that can decide the fate of Twitter. Sure, the headline is that the company will allow longer tweets to reverse its fortunes, but much more goes on under the hood.

Data science influences decisions large and small at Twitter. Aside from the new team, Twitter has roughly 30 data scientists – out of 3,800 employees – sitting in San Francisco.

“We are a small team. But by understanding our users, we can change the course of the company,” Linus tells me.

Sometimes there’s an art to it.

The new team will interpret data from the 80 percent of Twitterers residing outside the US, a group whom the company pins its growth aspirations on.

“I always tell candidates who are interviewing with us: this is probably one of the highest leveraged, most impactful roles you can have. Even if you are coming in as a junior data scientist, it’s not unheard of for data scientists to have direct access to the executives.”

Yet it won’t take much convincing for data scientists to join Twitter. With hundreds of millions of users worldwide, the social network springs forth an endless river of data – petabytes of them.

Data science isn’t always sexy

Special tools and processes are needed to handle the massive dataset. Here’s what Twitter’s data stack looks like:

A massive lab

The data science altar