Our R&D team is committed
to bleeding-edge research.
Editing images is a time-consuming task, is there a standard of beauty for pictures? Can an algorithm learn the best enhanced image for a given RAW photo?
The problem is clearly ill-posed and multiple solutions exist, as of today none has been able to produce a high-quality results.
Detection, segmentation
and image parsing
When editing an image, professional editors care about the context and apply different techniques based on the type of photography and the content of the photo. Creating AI for fully automated editing means knowing exactly what’s in the picture and where. With pixel accuracy, nothing less.
Learning beauty
If beauty is subjective, then the best picture for the given client is an optimization problem. If Netflix wants to know what movies you like, we want to know how you like your photos. If cold-start, domain-adaptation, and continuous learning are the words you’re thinking of right now, then we’re sending you a virtual high five!
Real-Time systems
Notifications: we are a time-sensitive business, thus all new events have to be promptly notified. How do we deliver hundreds of thousands of notifications per minute efficiently from multiple systems to multiple channels? Streaming image manipulation: photography is at the core of our business, thus high-resolution image processing is essential. How do we process millions of images per month? How do we build an infrastructure that scales seamlessly?
Manage Terabytes of Data
Image compression and reliable transfers: delivering high-quality photographic services means transferring gigabytes of data per second, how do we optimize bottlenecks even for slow connections? How do we make the process resilient and the user experience smooth? System monitoring: when data is transferred to our systems from all around the world 24/7, knowing what’s going on in our network is of paramount importance. How do we monitor a distributed infrastructure and reduce the MTTR?
Supply and demand adaptive matching
Our business leverages a wise balance between supply and demand. How do we find the best photographer for that specific photoshoot in that particular city of the world? Every market is volatile, how do we forecast the most accurate price for photoshoots around the world? How do we dynamically adjust it?
See more +
Head of Technology
Umberto di Fabrizio