Research and teaching in data management and data science and serving as Chair of the Department of Computer Science. I also play a leadership role in the campus-wide effort to grow Computer Science and Data Science at the University of Chicago.

As Director of the Berkeley AMPLab and one of the Original Creators of Apache Spark I have been a technical advisor and consultant to the company before, during, and since its founding.

Hyde Park Angels is the most active early-stage investor group in the Midwest. With a membership of over 100 successful entrepreneurs, executives, and venture capitalists, the group provides critical strategic expertise to entrepreneurs and the entrepreneurial community. More than half of our members have founded a company; most are CEOs, top executives, or corporate board members; and 100% invest in startups.

Research focus on query processing and advanced analytics in networked environments, clusters, and distributed computing clouds. At Berkeley I worked with fantastic students and colleagues to develop a number of systems such as the Berkeley Data Analytics Stack (BDAS) and (now, Apache) Spark for Big Data analytics, TelgraphCQ and HiFi for low-latency and distributed stream processing, and TinyDB for sensor networks (now called the Internet of Things). I taught undergraduate and graduate courses on database systems, data science (including the first Data Science course at Berkeley), operating systems, and distributed systems, as well as the occasional foray into Management of Technology. (Siebel Professorship Awarded 2012)From 2011-2016 I also served as a Director of the storied Algorithms, Machines and People Laboratory (AMPlab).

Primary responsibility was to help the amazing group of students, faculty and staff at this world-leading department keep doing what they do so well: pushing the envelope of Computer Science.

Working with Cisco managers and developers on long-term topics related to data analytics in general and the nearer-term integration of the Truviso technology into the Cisco Prime Analytics product line. Truviso's technology is aimed at "Data on the move" and so serves as a highly-efficient substrate for Network and Networked Analytics as well as for edge computing in IoT deployments.

Working with faculty and students who are launching this new research initiative in Shanghai. Also visiting and giving talks at a number of research centers, universities, and events around China.

After building the initial (world-class) team and serving as founding CEO, I was responsible for product and technology directions for this enterprise-oriented start up focusing on zero-latency data management in networked environments. Building on research on streaming query processing done at Berkeley, Truviso pioneered the idea of combined Stream-Relational processing in which analysis of live streaming data is combined seamlessly with historical data and previously-computed streamed results. This powerful concept enables not only the expected Real Time applications such as event processing and internet-of-things, but also turns out to be the ideal way to efficiently process the huge volumes of continuously arriving data in modern Big Data applications. Truviso technology is able to process up to a million new records per second on a single machine, and can scale horizontally when needed. Successful customer use cases include security, network analytics, clickstream and web analytics, and high-speed financial trading and analytics. Truviso was acquired by former customer Cisco Systems in 2012.

I spent part of my sabbatical in 2003 working with Mayfield partners and portfolio companies on a wide range of technology issues in data management. I also performed due diligence and research into business cases and ecosystems for what is now being called "The Internet of Things" or "Internet of Everything" or "Industrial Internet", etc. Attended partner meetings, company pitches and site visits. Learned a lot about start ups and Silicon Valley. A fantastic learning experience. Incidentally the rest of the sabbatical was spent visiting research groups in Paris and Hong Kong - very cool!

Research on distributed databases and data broadcast (including the "Broadcast Disks" system), early work on low-latency query processing and XML pub-sub that led towards stream query processing and event processing systems. Mentored some great Ph.D. students and postdocs. Enjoyed living in the DC area.

Development of highly-parallel database systems. We built a 40-server prototype of the Bubba Massively Parallel Processing (MPP) database system, which was designed with an (unheard of at the time) scalability target of 1000 nodes. Innovations from the project include: parallel query processing algorithms, micro-kernel architecture with a flexible communications library, memory-mapped storage with automatic kernel-level transaction support (my main project), deep integration with a functional programming language, and column-oriented storage. People looking at current scalable Big Data architectures would see a lot of things they recognize in this system from the 1980's, and maybe a few things that they haven't even thought of yet!

Development of database, information retrieval, and office systems. Developed several key technologies such as a highly-compressed index structure for sparse arrays (really handy for information retrieval and document search) and a client-server storage system to support collaborative applications in workgroups.