SINGAPORE — Twitter is searching for a team of data scientists to be based in Singapore, in its first effort to create such a team outside of the U.S.
Twitter's Linus Lee, who will be moving from Twitter's San Francisco headquarters to Singapore to head the team here, told Mashable the team will be between five and 10 strong, and will focus squarely on growing markets outside the U.S. — where 80% of the company's users now are.
Lee, a Singaporean who left the island state 10 years ago for undergraduate studies at Stanford, was one of Twitter's early data science hires four years ago.
In the time since he's joined, Twitter's data science teams have multiplied to encompass a business-focused team — that advises the finance side on trends — and other teams within engineering, although Lee wouldn't specify how many data scientists Twitter now has.
His own team in the U.S. works with the product development folks to shape the user experience, depending on what the data reveals about user behaviour. The new Singapore team will also be steered in this direction, and will focus on how users in emerging markets with less capable devices on slower networks act on the network.
For instance, in many of these markets such as India, users operate on 2.5G networks, which are far slower than the 4G speeds that mature markets are accustomed to. Based on user bounce rates shown by the data, the Twitter app now adjusts for slow connections and downgrades images or videos, to help the user experience along, Lee said.
"The data science team looks at the data, models it and recommends so the product team can test outcomes," he said "Previously, emerging markets teams had to beg for data science support, but now they'll have a dedicated resource (in Singapore)."
What Twitter is looking for
The job ad for the new Singapore team is already up on Twitter's site, and Lee said they're looking to hire both junior and senior analysts.
"I don't think age is a factor, but what separates people is the ability to apply the (data) tools and techniques in the right way. To understand what the business needs, and apply it.
"We need independent thinkers, because they need to suggest (features) to our product people," he said.
He noted that Singapore continues to produce more qualified statisticians out of schools here, so the scene is ripe for hiring. But many here will be inexperienced compared with counterparts in Silicon Valley when it comes to handling huge data sets, because there are fewer opportunities here to handle data the scale of what Twitter has.
He said Twitter now has 310 million monthly actives, and data from 500 million logged-out visitors per month. Hundreds of millions of tweets are sent daily, so the vast amount of unstructured data gets crunched continuously by the social media giant.
Tens of thousands of Hadoop jobs run overnight in batches across Twitter's data centres holding petabytes of information, he said.
This effort to chase new users comes as the company struggles to keep user growth up for investors. The topic has been a pain in its side in recent years, and the firm has started to tinker more with the fundamental experience in hopes of jumpstarting a stalled user base.
This includes algorithmically tweaked timelines — which, predictably, drew irefrom users — and considerations of lifting the 140-character limit, which has long been considered the hallmark of the service's appeal.
"We seek to understand different users," Lee explained. Users in Japan, for instance, mostly choose to be anonymous on Twitter, whereas those in the West prefer to tie their real-life identities to the service.
"In the last four years, the proportion of users outside the U.S. has only been growing. You can't just apply what you know (about the U.S. base) over here," he said.
Have something to add to this story? Share it in the comments
No comments:
Post a Comment