I don't know your case in details. Do you want to only detect faces on the images, or verify the identity of a person using the face? Or maybe you want to implement something more sophisticated? In any way, I would research the solutions from the top cloud providers first. You can compare supported features, prices, and convenience of usage and then decide what is your best option. IBM has its Watson Visual Recognition service, Microsoft Azure has the Face API as the component of the Cognitive Services, Amazon Rekognition is responsible for computer vision in the AWS platform, Cloud Vision API is the similar product of Google. It's up to your choice.