As Chief Technology Officer at Zefr, the global leader in brand suitability targeting and measurement across walled garden environments, I am responsible for the company's strategy, formulation, development and cross-functional delivery of our technology products. Creation and growth of SAAS measurement line of business with 100% YoY growth at 10M ARR. Known for continuous innovation, my team builds Zefr's platform through a combination of advanced machine learning, AI and data science.

Leading an amazing team that builds a cross platform video analysis data platform for the largest walled gardens.

Making Youtube ads less annoying for users and more palatable for advertisers through the magic of machine learning and content relevancy.

Led the Front End, Back End, Data and SRE teams to the only data science backed, video level content relevancy solution for Youtube videos. Successfully focused the team on creating a new product focused on the programmatic space.

• Developed and maintained front and back end infrastructure• Built up audience using SEO and social media• More than 18M lifetime users.• Top 50,000 website worldwide

• Unified high performance event based architecture with a high performance message bus, standardized message formats, format discovery, data lake archival.• Leading technical migration from an agency model to a SAAS based model with public APIs and websites.• Designed and led team to implement end to end machine learning video classification platform• Directed Google Adwords automation project, enabling near real time video level targeting• Led 2 teams of 4 and 6 people

• Rebuilt platform engineering team into cross-functional, highly regarded team, with 4 direct reports.• Designed and led 12 person cross-functional team implementing a public API and authentication architecture, improving response times on our highest traffic pages.• Redesigned photo infrastructure and implemented a CDN after an extensive cloud based evaluation of different vendors.• Designed and led team implementing a send time optimized recommendation notification system.• Expanded number of nightly social recommendations by 4X to 10 million, and speed by 2X with a redesigned social recommendation big data engine. The system was eventually scaled to 25 million daily recommendations through horizontal scaling.. • Designed a new service oriented architecture and Hadoop batch processes by leveraging horizontally scalable document and key/value stores. • Raised member communication rate 2X by introducing sophisticated recommendation and user analytics models. • Reduced latencies of real time recommendation system by 10X with a high speed, parallel processing scoring service powered by a big data model training infrastructure.• Architected the horizontal scaling of key services to withstand a 4X increase in site traffic.

• Improved accuracy of social recommendations through real-time aggregating, analyzing, and compressing user communication data for use in models.• Lowered model re-training time by creating a generic scoring library with dynamic code generation.• Sped up various Hadoop processes by 2-10X through map side joins, remote key/value joins, combiners and job tuning.• Increased network transfer performance by implementing REST server using protocol buffers.• Saved $150K/yr. and increased reliability by migrating EC2 to an internal Hadoop cluster.

• Led the pre-trade product portfolio team to serve 150K requests for 10M calculations daily.• Designed and implemented a historical stock analytic ftp service of over 1 million files.• Overhauled website to leverage a dynamic portlet framework with web service integration.• Developed clustered caching for multi terabyte Sybase database.• Automated on demand Excel report generation with complex graphs and charts for reports.• Improved monitoring system batch processes and server components to early alert time critical deadlines in danger of being missed.