CLIENTS & PROJECTS
Over the years, we have provided many custom solutions to dozens of clients in various stages of their business, from pre-seed to traditional industries. Here are some of the projects we can share:
D-ID is a Y-Combinator company revolutionizing the world of image & video protection from facial recognition. They use cutting edge deep learning technology to protect the identity of faces from being misused. We worked with them to help support this ability by developing models that use general adversarial networks (GANs).
Roofr is a Y-Combinator backed startup that modernizes the roofing industry. We helped Roofr by building a proprietary machine learning algorithm to segment roofs with pixel perfect accuracy from aerial imagery, and classify roof types and conditions using convolutional neural nets.
Retail AI are bringing state-of-the-art computer vision to bear on brick and mortar retail spaces. We worked on training deep learning convolutional networks on video footage from grocery stores for fraud identification, action recognition, and scan counting. We built custom solutions from the ground up to solve novel problem definitions with multiple objectives and losses.
Sown to Grow is an Ed-Tech startup that helps students in the U.S by promoting self-reflection in students. We built an NLP model to aid teachers in scoring student essays for whether they contain insight, a concrete plan to action, and other parameters that are empirically shown to promote success in student outcomes.
Skyline AI, a company making waves in the world of real estate, is a Sequoia backed venture bringing artificial intelligence into the world of commercial real estate investing. Having raised over 20M, they are committed to bringing the latest technologies to bear on the market, and we have helped them with problems such as sales estimation and rent forecasting using deep learning techniques.
L1ight develop a suite of solutions to combat online toxicity, using machine learning. We worked with them on their anti-toxicity platform, building models to identify suspicious behavior from text and images using advanced NLP and vision techniques.
Started as a B2C photo curation startup, we helped RealFace build their own face recognition solution using groundbreaking deep learning techniques.
The face recognition solution we helped to build went on to become the most valuable aspect of the company, which renamed to RealFace and ultimately sold to Apple, helping power the face recognition on every iPhone on the planet.
SecuredTouch are innovators in the world of behavioral biometrics for mobile, delivering continuous authentication technologies to strengthen security and reduce fraud. We helped them improve their capabilities by building a machine learning model to identify whether or not a mobile user was perpetrating fraud by comparing their gestures to their profile.
Ridevision are at the intersection of driving and technology, bringing desperately needed safety technology to the world of motorcycles. We built a POC for them to apply deep learning to motorway video footage to predict the time-to-collision with oncoming vehicles, to help develop technology that will alert drivers before an impending accident occurs.
Climacell is a series-C startup that revolutionizes weather forecasting by tapping in to unorthodox data sources to better understand the weather. We helped Climacell build their first visual deep learning model to infer weather conditions from video streams, owning the entire pipeline from data collection & annotation to setting up a scalable production system that analyzes video streams across the US for an up to the second visual snapshot of the weather.