Microsoft Corporation (2017 - Present)
I am a Principal Group Program Manager in the Azure AI Platform group, where I lead product development, technology planning, and business development strategy for Azure Cognitive Services.
Intel Corporation (2016 - 2017)
I was the Chief Product Officer at Intel® Saffron™ Cognitive Solutions Group in Hillsboro, Oregon, where I was responsible for product engineering, product marketing, product management, and product development of enterprise-ready AI solutions. I was previously an Entrepreneur-in-Residence (EIR) and Research Science Manager in the New Devices Group within Intel's New Technology Group. My areas of interest included artificial intelligence, machine learning (including deep learning, of course), natural language processing, knowledge representation and reasoning, and computer vision.
Vulcan Inc. (2014 - 2016)
I co-founded (along with Keith Rosema and Alan Caplan) the Vulcan Proving Ground, an internal incubator dedicated to growing Vulcan's strategic technology portfolio.
My focus areas included mobile devices and wearables, machine learning, artificial intelligence, and computer vision. I also worked with teams interested in unmanned aerial vehicles, immersive environments (AR/VR), brain injury science, and molecular diagnostics. I served as CTO; I was responsible for technology scouting and product development. I also spent a year working closely with Microsoft co-founder Paul G. Allen to create a technology investment roadmap that bridged his philanthropic and strategic investment goals. While in that capacity, I helped organize Vulcan Inc.’s technology response to the 2014 Ebola crisis in West Africa in conjunction with the United Nations, USAID, CDC, and multiple NGOs.
A.R.O. Inc. (2010 - 2014)
A.R.O. Inc. was a startup that I co-founded with Paul Allen in Seattle. We developed software that produced insights about customers' offline behavior from sensor data captured by their mobile phones. Our flagship app, Saga, was a lifelogging tool that more than 5 million people downloaded. You can read more about what people said about Saga here.
Swingly (2008 - 2010)
Swingly was a Web scale question-answering engine that found exact answers to factual questions. Development started in Spring 2008, and our beta launch occurred in August 2010. Swingly was based on a simple idea: what if you could turn the Web into the world's largest FAQ? We believed that if you could index specific answers (and not just documents), you could help people find the information they were looking for more quickly and easily than with traditional search.
Swingly was built by a small team (including Ying Shi, Tobias Jungen, Jeremy Bensley, John Williams, and me) -- but benefited immeasurably from the excellent natural language processing and natural language understanding software developed at Language Computer Corporation.
Extractiv (2008 - 2010)
Extractiv was a Web-based content-as-a-service startup started jointly between Language Computer Corporation and Creeris Ventures. Extractiv allowed users to transform unstructured Web content into structured semantic data -- from their desktops. It combined a powerful Web crawler infrastructure (made available by 80Legs) with the amazing text extraction software developed by LCC. The result? a Web service that could crawl, download, and mark-up unstructured text faster (and with better precision) than other data center-based solutions. It was cheap, too!
Extractiv was the second natural language processing company incubated at Language Computer Corporation during my tenure as CEO. John Lehmann and Shion Deysarkar led the company from its founding in 2009 to its closure in 2011.
Language Computer Corporation (2004 - 2010)
Language Computer Corporation (LCC) develops cutting-edge natural language processing software solutions for operational customers in the U.S. Defense and Intelligence Communities. I worked at LCC from 2004 - 2010. I started out leading research and in 2008, became CEO. I was so fortunate to work closely with Dr. Sanda Harabagiu, who was a real role model of mine in so many different ways.
During that time, I ran a laboratory committed to the acquisition of high-quality knowledge from large amounts of textual information in any language. We invented systems for information extraction, question answering, textual entailment, and document summarization that were judged as some of the world's best in evaluations like TREC, TAC, DUC, and PASCAL RTE.