Software Engineer, Personalization and Platforms @ During Yahoo’s rebuilding period, I joined the Ad-Targeting group as a Platform Engineer focused on near real-time processing, moving 280k peak events per second. Before going real-time, the scoring pipeline was run on the Hadoop infrastructure, taking up to a week for events to propagate. Over the course of a year, I worked on key Storm components with my team lead, and several team members from Beijing, to reduce the latency of the pipeline from several days to within seconds. The net result was an improved ad experience for users, by updating their profile in near real-time. As part of redesigning the pipeline, I delivered a new component that transformed and loaded data onto memcached servers read by Storm. We decided to make it a framework, and met that goal by abstracting the transformations from the system code, allowing new development to occur by simply adding a single Java POJO. Using Google’s Guice library, the entire component was testable in-memory and enabled test-driven development. The component was setup for continuous delivery to meet a corporate goal. From June 2013 to February 2015 (1 year 9 months) Staff Engineer @ Baynote is a SaaS product providing e-commerce product recommendations. We had a proprietary platform that was built before Hadoop went mainstream, and I worked aggressively to fix many concurrency and performance issues my first year. We later rebuilt the product on Hadoop to meet scaling demands, and used Hive as the central processing tool. I dug into the internals of Hive to learn its serialization mechanisms, and learned about the Hadoop ecosystem in general. I was given the task of creating a new workflow manager, and decided to use Spring Batch as a basis. Over the course of 18 months, our delivery of the new platform enabled us to process much more data, attracting customers with higher scaling needs. Baynote was a small and growing company when I joined. I interviewed over 30 candidates for both Engineering and Product Management positions in a couple of years. I demonstrated expert level knowledge of our old platform during the transition period onto our new platform, and was instrumental in propagating system knowledge to the growing team. We built a team culture that recognized us as a top place to work by The Bay Area News Group in 2014. From January 2013 to May 2013 (5 months) Software Engineer @ From August 2010 to December 2012 (2 years 5 months) Software Developer @ I returned to BayTSP full time preceding my senior year, and took the fall quarter off. I implemented new crawling technology for the company to target sites without boolean search operators, and created new crawler profiles using our platform. Alongside John Bernardo, we delivered a new BI Dashboard that was revealed to our customers at the company’s annual piracy conference. Under direction of the VP of Research, we built a prototype ad technology that delivers contextual advertisements based on video content. From July 2009 to March 2010 (9 months) Lead Internet Analyst @ BayTSP was an anti-piracy company that served the largest names in media and software. As an analyst, I joined the team that gathered infringements and removed content online. We employed web crawlers to index the contents of piracy sites for review. However, once a site experienced hypergrowth, we were unable to index more than a fraction of its content. BayTSP considered using keyword based searches, but result limits restricted our view to the same capacity. This was a major problem impacting the business. To solve this, I recognized that we can split the results into two sets by exploiting boolean logic on the search term, and recursing through the search space. I doubled the results on each iteration and was able to extract millions of results for a target asset. I implemented the solution in C# and designed the output to work with our existing tool chain. The net result was that we discovered infringements 6X faster with fewer team members. With company permission, I wrote a technical paper called Recursive Descent Queries that satisfied my undergraduate writing requirement based on this technique. I was promoted to lead analyst based on this work, shortly after joining the company. From July 2007 to January 2008 (7 months)

BS, Computer Science, Cum Laude with Honors @ University of California, Santa Cruz From 2007 to 2010 Joe Napolitano is skilled in: Unix, REST, Distributed Systems, MySQL, Java, Agile Methodologies, Hadoop, Cloud Computing, Algorithms, Scalability, Artificial Intelligence, Software Development, Linux, Software Engineering, Spring