CrowdFlower, a company that helps customers build AI systems by providing them with training data, announced today that it’s getting into the business of helping companies implement machine learning.
While many companies are trying to incorporate AI and machine learning to further their business, there’s a dearth of talent out there to help them get started. That’s where CrowdFlower wants to come in, by providing support to get companies moving along the right path.
“We’re constantly being asked by our customers to give more guidance and play that trusted mentor role,” CrowdFlower CEO Robin Bordoli said. “[Companies will say] well, we’re thinking about the overall project, can you give us guidance on the overall project, not necessarily just the training data piece?'”
The addition of consulting could serve CrowdFlower costumers in a few ways. It could help existing customers get unstuck with a system that isn’t working, assist businesses that have already implemented one machine learning system get started with something completely new, and also get brand new customers started with implementing AI.
The company announced today that it hired Robert Munro as its vice president of machine learning to spearhead those efforts. Prior to his work at CrowdFlower, Munro was a principal product manager working on Amazon Web Services’ natural language processing and machine translation products.
He’s also a longtime CrowdFlower user. Munro started using the company’s services in 2010, when he was working on the U.S. government’s response to a massive earthquake that struck Haiti.
“Working with 2,000 members of the Haitian diaspora, we used the CrowdFlower platform to structure information, which we could also then apply AI to, so we could pull out the most important and time-critical pieces of information,” he said. “This was something that we were able to launch in just 48 hours.”
This doesn’t mean that CrowdFlower is abandoning its work providing companies with training data — quite the contrary. The consulting push is more of a move to augment its existing business, Bordoli said in an interview.
CrowdFlower remains committed to its existing platform, which allows companies to submit data for labeling by a crowd of certified humans. That labeled data is important for the creation of supervised machine learning systems, which make up a great deal of the AI applications people are working on implementing in business today.
In addition to the consulting news, CrowdFlower also added two members to its Scientific Advisory Board. Monica Rogati and Adrian Weller — both veterans of the machine learning ecosystem — will give CrowdFlower input on its technology and product strategy, as well as advising the company on developments in the AI ecosystem at large.