Join Some of the Brightest Minds in AI
Turn cutting-edge ML techniques into enterprise-ready AI solutions for the world's largest organizations
I joined Hyperscience in April 2020, having spent the previous eight years at Google, most recently as the head of the Perception AI team responsible for their computer vision and AutoML developer platforms. I was immediately struck by the opportunity becoming VP of Product & Design at Hyperscience presented: it was a unique chance to lead and grow a talented Product & Design organization, work with cutting-edge AI, and transform how enterprises leverage automation to better serve their customers.
As I reflect on my time with Hyperscience thus far, I am extremely grateful for this opportunity. In 2020 we grew almost 3x as a company, and as momentous as last year was for our organization, I am more excited about what we’re building in 2021 and beyond.
In the last few years, Hyperscience has evolved from an Intelligent Document Processing startup into a global company with 250 employees, automating workflows for dozens of Global 2000 enterprises and government agencies, and in turn, impacting the lives of millions of their customers. Thanks to our technology, our customers are able to approve insurance claims faster, pay their vendors’ invoices on time, and help citizens get quicker reimbursements from their health insurance.
It’s magical to see how Hyperscience has evolved and elevated document understanding AI from experiments in research labs into enterprise-grade solutions that deliver real outcomes for real companies and real people. Something I couldn’t have dreamed of a decade ago.
But this is just the beginning.
We need to help enterprises do even smarter work, empowering them to deliver valuable and fair outcomes for all their end customers. Manual work is slow, expensive and error-prone, and data entry errors lead to poor decisions and outcomes with real societal impact.
The mortgage application process is one example of this. It costs tens of thousands of dollars for an originator to process a single mortgage because of the need to manually process the hundreds of pages of documentation that come in a mortgage packet. And the final outcome is almost surely suboptimal because loan processors and underwriters have limited or skewed data about their prospective borrowers.
This inability to scale manual processing – and the need for rules-based processing – has led to billions of human hours going to waste doing repetitive keying work. It has also resulted in a disparity of treatment, as exemplified by the redlining process of the 1900s, whereby U.S. lenders used checklists and zip codes to assess risk. This led to people with equal solvency but different zip codes receiving different approval rates and caused a major impact in access to housing and economic development for working-class communities.
With automation, we have an opportunity to fix this, and that is one key goal of our platform.
In June, we shared our vision for the future, which we’re calling Software-Defined Management, or SDM.
At the center of this vision- and what makes Hyperscience so unique – is our state-of-the-art AI, which is tightly integrated with our proprietary human-in-the-loop user experience. Our models are exceptionally good at knowing when they’re going to be right, and when they’re likely to be wrong, bringing in humans to review and resolve edge cases in a bite-sized, intuitive fashion. With Hyperscience today, a data keyer who used to do tedious, repetitive entry can now rely on our platform to assist them. Plus, the more the system is used, the better we get and the more we are able to help them out moving forward.
Software is becoming intelligent, and with Hyperscience’s proprietary technology, our customers’ business processes can learn and become more efficient and effective over time, starting with data keyer work and evolving to automate all forms of repetitive knowledge work.
The next generation of our platform will evolve our leading IDP solution into the world’s first input-to-outcome automation platform. This will make increasingly unstructured data formats (e.g., emails, digital documents, images) machine-readable, building a semantic understanding of data by mapping them to a universal taxonomy of business objects. Using Hyperscience, businesses will be able to snap together horizontal, stackable blocks and workflows to build vertical solutions that automate key business processes such as claims processing or loans origination, seamlessly involving humans in the process to render a faster and more informed decision.
Our Product, Design & Engineering team is recruiting across our New York City, Sofia, London, Toronto, and remote locations. We’re looking for imaginative, collaborative individuals who are interested in solving exciting, technical problems (such as multimodal AI, knowledge representation, human-in-the-loop interfaces, dynamic user interfaces and self-learning platforms), all while stretching personally and professionally.
How to Build and Structure Great Product Teams via Built In NYC