Research and Technology
Building an (artificially) intelligent solution is as much art as science. Building a product out of it is another story altogether.
The problems we deal with are non-trivial; and we use a hybrid stack of complex and proprietary image processing algorithms, in conjunction with classical machine learning techniques. There's a lot of deep learning usage in some of our modules as well. We are deeply grateful to the open source libraries and sometimes we contribute to these projects to give back.
Productizing a system with so many layers of complexity is an effort that needs to talk to APIs on one side and customers on the other. It needs nuanced thinking, discipline and flamboyance in equal measure.
It's challenging, but if you're up for it, let us know.
October 10, 2017