I’m passionate about creating scalable and efficient software for the toughest challenges in distributed systems. I enjoy participating in all phases of software engineering, especially system evaluation, architecture design, and deployment. I strive to automate everything and be involved with all parts of systems including production support. Specialties: • The Hadoop stack: HDFS, MapReduce, Hive, Luigi, Kafka, Zookeeper. • High-performance database and distributed caching systems: Memcached, MySQL, HBase • Java & Scala • Linux system administration, deployment, and automation using Puppet and Chef

Software Engineer @ From September 2015 to Present (4 months) Washington D.C. Metro AreaSr. Software Engineer @ Working on all things backend. From September 2013 to May 2015 (1 year 9 months) Greater New York City AreaSoftware Engineer @ working with big data. From January 2012 to August 2013 (1 year 8 months) Greater New York City AreaTechnical Staff / Senior Software Engineer @ • Advocated for, designed, and lead member of a small team that implemented Adconion’s Hadoop-based analytics infrastructure. • Devops – authored modules for Hadoop stack deployment and configuration using Puppet. • Continuous Integration – all elements of software stack, including pig-based workflows. • 2010 Excellence Award selected by peers as among the top-5% of employees in the company. • Adconion acquired Joost in November of 2009. Learned and personally managed all backend services for Joost.com’s relaunch in early 2010. From November 2009 to December 2011 (2 years 2 months) Greater New York City AreaMember of Technical Staff @ • Designed and implemented Java backend APIs for Joost’s social networking and online-video features as part of a three-person team. Authored over 30K lines of code for a service that scaled to over 3 million monthly unique users and thousands of REST API users. Technologies used include Ant, Spring, Junit, Lucene, Jetty, MySQL, Memcached, JavaScript. • Involved in product management cycle, as an advisor and authoring requirements documents. • Implemented internal tools for editorial management of promotions on the website using the Google Web Toolkit (GWT). • Implemented prototype Content Management System using Django and Python. • Co-developer of Joost’s Facebook Connect implementation, which was the most-used implementation of Facebook Connect for months after launch From March 2008 to November 2009 (1 year 9 months) Research Assistant @ • Authored a medium-sized (10,000+ lines) C++ software package to partition massive graphs (more than hundreds of millions of edges). The software is designed to scale for graphs that are too large to fit in RAM. • Created utilities in PHP and Perl to form a graph from the results of a MySQL query. • Implemented C++ services to answer graph queries. From January 2007 to February 2008 (1 year 2 months) Teaching Assistant @ • Led recitation on elementary algorithms and data structures in Java for a recitation of 25-50 students. • Created project assignments, held office hours for students, and graded exams. From September 2006 to January 2008 (1 year 5 months) Software Developer @ • Created an interactive Flash-based web-program to measure implicit attitudes using response latency. Used PHP and MySQL for data storage. • Developed web-based survey using PHP and JavaScript. • Software has been used for 5 studies across thousands of participants. From September 2006 to November 2007 (1 year 3 months)

MS, Computer Science @ Rutgers University-New Brunswick From 2006 to 2010 BS, Computer Science @ Lafayette College From 2002 to 2006 Joe Crobak is skilled in: Hadoop, Java, Python, Lucene, Memcached, Django, Dojo, Spring, JUnit, Ant, Maven, GWT, Automation, Puppet, EC2