To boost its in-house, machine-learning efforts, Twitter has acquired Whetlab, a startup that makes it easier for companies to use machine-learning tools.
As part of the acquisition, announced Wednesday, Twitter will shut down Whetlab's beta service on July 15, and will no longer accept sign-ups for the product. Current users will be able to export their data from Whetlabs's website in either tab-separated format or JSON.
It's not clear how Twitter plans to use Whetlab's technology to enhance its existing machine-learning plans. However, the startup's tool seems useful for any company implementing machine-learning techniques. The technology, which was developed by researchers at Harvard, Toronto and Sherbrooke universities, takes in information about the problem a user wants to solve with machine learning. It then gives the user a series of suggestions to help them optimize a machine-learning model to solve the problem.
Users don't have to send Whetlab their data, nor do they need to use a special machine-learning toolkit, according to the company's product page. "You can run your code on your own private machines at whatever scale you want. What we do is help you use those resources optimally by telling you what you should try next," the page reads.
The acquisition makes sense for Twitter, which is trying to offer users more customized experiences to improve adoption of its microblogging service. Machine learning is likely critical to those initiatives, and acquiring technology that makes the application of machine-learning technology easier could benefit Twitter's in-house development process.
The deal comes at a tumultuous time for the company, which announced last week that CEO Dick Costolo would be stepping down on July 1.