GattaccaNotesHome!

About

avatar

Miles Zarn

Data
MyDataWiz
mailgithublinkedintwitter

Abstract

Miles Zarn 20+ year professional in Data, Data Platforms, and Data Engineering. Hobbies include hacking node, next, tailwinds, and data modeling with data catalogs and dictionaries.

Miles has worked on massive 50k node data platforms storing petabytes of data, developing and leading phased deployment, experiments, and behavioral data analysis platforms and insights.

Experience

ASAPP

Director of Engineering, Data Platform & Analytics
ASAPP
2019 – Present
San Francisco Bay Area

  • Retain existing team and grow team by 100%. Create team processes, controls, and knowledge sharing to enable distribution of architecture & OnCall activities from single individual to the entire team. Create actionable operational Dashboards, Alerts and RunBooks. Significant measurable improvements in availability, reliability, scalability. Develop and evangelize 6+ month roadmaps, establish KPIs, contribute to organizational KPI development. Contribute to future architecture.

  • Create Quarterly Team Plans for review evangelizing & team capacity planning. Create prioritization and triage framework to manage 15 internal teams, 3 customer facing teams, and 6 customers data, analytics and related requests.

Walmart Labs

Director of Engineering, Behavioral Data

Walmart Labs

2018 – 2019

Sunnyvale, California, United States

  • Retain existing team and grow team by 50%. Take ownership and revive an imploding Platform team. Create team process, controls, and knowledge sharing to enable distribution of operational activities from single individual to the entire team. Create actionable operational dashboards, Alerts and RunBooks. Significant measurable improvements in availability, reliability, scalability. Improve USGM (largest vertical) by .29% gross revenue via code optimization and increased site speed. Develop and evangelize 12+ month roadmaps. Support critical instrumentation activities for marketing, advertising, affiliates, and category specialists.

  • Architect transformation from private to public cloud behavioral data ingest system and related data lakes to ingest 270k RPS or 8.6 GB/s behavioral data leveraging existing ecosystem and capability.

Credit Karma

Director Of Engineering

Credit Karma

2016 – 2017

San Francisco Bay Area

  • Recruit and retain a team of top engineering talent to increase data quality, scale, reliability and reduce latency of ETL and Event based data systems. Grow the team from 3 people to 17 ICs, and two managers within 4 months.

  • Migrate the entire company from Vertica to Big Query, reducing recurring costs, and query latency.

  • Reduce data reliability escalations from nearly daily occurrences to once every three months. Created data SLAs, and consistently increased reliability, scalability, while reducing latency, delivering predictability to the rest of the company. Create a pre-production Client Event QA environment to further increase event quality and consistency, addressing a critical need across VPs.

  • Created Google Cloud Service data warehouse for adhoc batch data mining, utilizing on demand Spark clusters. Created streaming Spark services to support data validation, remediation, hydration, transformation for various near real time data consumers, reducing latency, and engineering complexity, increasing engineering agility.

  • Partner with BI, and Analytics teams to collect the right data, on quality, and manage dependencies for reporting engines. Reduced data complexity by consolidating 26 different schemas for click and impression data to one, across all platforms.

Sony Network Entertainment Int'l (SNEI)

Sony Network Entertainment Int'l (SNEI)

Engineering Manager

2013 – 2016

San Francisco Bay Area

  • As an Engineering Manager, I identified, and evangelize the need to Exec staff to define, fund, and found a real time Big Data insights and Experiment Platform for the top 12 SNEI Client Applications (Game and Video Store applications) impacting a combined 4+ billion annual business.

  • Develop a patent pending phased deployment and experiment platform used by SNEI to reduce risk of deployment and generate statistically valid A/B experiment results. Also developed a patent pending client instrumentation tooling to provide real-time analytics for embedded, Web, Mobile, and i* applications.

  • Partner with global BI teams and client dev teams, to foster and enable data driven decisions for every feature shipped.

  • Recruit and retain a team top engineering talent to re-invent, scale, develop, support various aspects of a BIS data pipeline, to acquire, share, and visualize customer behavioral data systems providing real-time analytics, insights for online services, and actionable BI data using Databricks, Spark, Hadoop, Kafka, AWS Cloud technologies.

  • Participate in definition of SNEI core employee values to define corporate culture and employee performance goals and attitudes.

  • Identified, evangelized industry trends and technology by recruiting qualified 3rd party keynote speakers and started 4 major initiatives (User Error, User Latency, A/B testing and User Personalization) focusing client team efforts for 3 years using data driven decisions contributing to building the business from 1 Billion to 4+ Billion revenue.

Sr Technical Program Manager

2012 – 2013

San Francisco Bay Area=

  • Technical Product Manager in SNEI responsible to coordinate a team of 8 developers and broker cross team, cross division relationships to ship a client for Sony TVs, and Blue Ray disc players providing a shopping and viewing experience for movies and TV series content. Project completed on time and on schedule realizing many features that were not original planned.

  • Introduce, Create and Define Agile Program Management for 3 dev teams in San Jose office, providing predictable software delivery, even when working through significant unknowns. Create and hold meetings every 6 weeks to demo achievements, socialize difficulties, set schedule expectations, and identify critical path, and related dev partner goals. The method was so successful the San Francisco and San Diego offices converted to Scaled Agile Framework 9 months later affecting 300 developers.

eBay

Program Manager

eBay

2011 – 2012

Campbell, CA

  • Technical Product Manager in eBay responsible for Hadoop, Data Mining and Infrastructure, Application Monitoring.

  • Produced core layers of cloud automation and related monitoring infrastructure to support 2000+ servers and 20TiB/day data.

  • Interview customers; define durable program themes, impact, problem statements, success metrics and scope of work to focus engineering effort on high impact deliverables.

  • Enable catalog comparison reporting for motors with web crawling and data mining, identifying 20+ million dollars in potential revenue.

  • Manage site wide Monitoring and produce instrumentation, monitoring and alerting to critical behavioral data pipeline.

  • Deliver core functionality for eBay’s Cloud Automation solution scaling a distributed lock system 100x, Create an accounting scheme to allocate and secure could resources

Microsoft

Microsoft

Program Manager

2009 – 2011

  • Build the vision and roadmap to satisfy availability, reliability and performance requirements of ~ 50 business units demanding more capabilities for Cosmos, Microsoft’s version of a 200 petabyte, 50,000 node cluster of Hadoop like technology.

  • Realize 30% cost efficiency gain and various other crucial feature, saving nearly $7 million a month in operational expense.

  • Migrate services from deprecated datacenter, hardware and OS to new datacenter, hardware and OS while in production by creating a hardware, partner and data migration strategy that required planning and coordinating of hardware provisioning, OS testing, Software system limit testing, customer and data pipeline migration.

  • Reduce job runtime 6% through a batch operation strategy developed by analyzing jobs, identifying operations that can be grouped, planning and coordinating experiments, managing customer trials and identifying success criteria.

Microsoft Research, Software Engineer

2007 – 2009

  • Research Software developer on Bing research team responsible for improving search relevance and customer experience.

  • Own the goal, metrics, process and feature work in a dynamic development environment.

  • 4% reduction in customer click dissatisfaction, improving Bing relevance using search click logs analysis to identify and remove low value spam, made for ads and SEO pages.

  • Improved Search results through a focus on the largest problem strategy by automatically debugging customer dissatisfied queries providing clear problem areas to focus on and monitoring progress.

  • Own the vision, roadmap and define scenarios and requirements for automatic relevance diagnostic reporting and supporting the related investigation needs Bing feature teams.

Symantec Software

Symantec Software

Principle Software Engineer

2001 – 2007

  • Principal software developer for VERITAS Storage products responsible for Hardware Compatibility and the install and upgrade experience for two suites of products.

  • Wrote the first ever multi-platform Storage Foundation Hardware Compatibility List (HCL) tool. Defined and created APIs to reduce test case code. Created a test feedback system to track test harness and test suite quality and effectiveness build over build. Coded 20% of the test cases requiring in-depth knowledge of: RAID arrays, SANs, IO subsystems, and products on 7 platforms. Supported internal and external customers to qualify thousands of configurations from tens of vendors supporting $400 million per year business.

  • Influence 13 product teams in 5 global locations to create a homogeneous install and upgrade experience on 7 platforms. Defined APIs and packaging rules to support individual product install and two suite installers. Played a crucial role in consistent use of authentication and authorization components across all offerings.

  • Wrote install code, managed daily install quality and debugged product problems for Storage Foundation and CommandCentral product suites. This role required knowledge of RAID arrays, SANs, IO subsystems and products on multiple operating systems.

Gadzoox

Customer Engineering Manager

Gadzoox

2000 – 2001

  • Owned hardware compatibility testing of OEM configurations. Collaborated with sales to define customer requirements and hardware compatibility matrices. Executed testing critical to winning new and retaining existing OEM contracts.
  • Responsible for switch performance testing. Coordinated with hardware and firmware engineers to formulate and execute experiments to drive the performance boundaries of ASIC designs.
  • Managed and trained a team of 5 junior engineers. Trained individual members to setup and execute Fiber Channel and RAID array tests, analyze protocol exchanges and debug hardware. This enabled the team to operate efficiently in a highly technical fast-paced environment.

Force Computers

System Architect

Force Computers

1998 – 2000

  • Owned hardware compatibility testing of OEM configurations. Collaborated with sales to define customer requirements and hardware compatibility matrices. Executed testing critical to winning new and retaining existing OEM contracts.
  • Responsible for switch performance testing. Coordinated with hardware and firmware engineers to formulate and execute experiments to drive the performance boundaries of ASIC designs.
  • Managed and trained a team of 5 junior engineers. Trained individual members to setup and execute Fiber Channel and RAID array tests, analyze protocol exchanges and debug hardware. This enabled the team to operate efficiently in a highly technical fast-paced environment.

Patents

System and method for remote monitoring of API performance and user behavior associated with user interface

  • Patent number: 10110688
  • Abstract: Various aspects of a system and a method to remotely monitor API performance and user behavior associated with a user interface (UI) are disclosed herein. In accordance with an embodiment, the system includes a server that includes registration of a metric associated with performance of an application program interface (API) and/or the UI associated with a client device. The metric may be registered as a unique namespace information value. A numeric value is assigned to the registered metric to associate the assigned numeric value with the registered metric. The assigned numeric value associated with the registered metric is synchronized at the client device and the server. The performance of the API and/or the UI associated with the client device is determined based on receipt of the assigned numeric value from the client device.
  • Type: Grant
  • Filed: March 17, 2016
  • Date of Patent: October 23, 2018
  • Assignee: SONY INTERACTIVE ENTERTAINMENT LLC
  • Inventor: Miles Zarn

Method and system for processing a geographical internet protocol (IP) lookup request

  • Patent number: 10075411
  • Abstract: Various aspects of a method and a system to process a geographical internet protocol (IP) lookup request of an IP address are disclosed herein. The method is implemented in a first server. The first server receives a first request for the lookup of the IP address from a first computing device. In response to the received first request, data is retrieved from a cache memory. The retrieved data from the cache memory is communicated to the first computing device. The cache memory stores the data received in response to a prior received request for the lookup of the IP address.
  • Type: Grant
  • Filed: January 7, 2015
  • Date of Patent: September 11, 2018
  • Assignees: SONY CORPORATION, SONY NETWORK ENTERTAINMENT INTERNATIONAL LLC
  • Inventor: Miles Zarn

SYSTEM AND METHOD FOR REMOTE MONITORING OF API PERFORMANCE AND USER BEHAVIOR ASSOCIATED WITH USER INTERFACE

  • Publication number: 20160277264
  • Abstract: Various aspects of a system and a method to remotely monitor API performance and user behavior associated with a user interface (UI) are disclosed herein. In accordance with an embodiment, the system includes a server that includes registration of a metric associated with performance of an application program interface (API) and/or the UI associated with a client device. The metric may be registered as a unique namespace information value. A numeric value is assigned to the registered metric to associate the assigned numeric value with the registered metric. The assigned numeric value associated with the registered metric is synchronized at the client device and the server. The performance of the API and/or the UI associated with the client device is determined based on receipt of the assigned numeric value from the client device.
  • Type: Application
  • Filed: March 17, 2016
  • Publication date: September 22, 2016
  • Inventor: MILES ZARN

METHOD AND SYSTEM FOR PROCESSING A GEOGRAPHICAL INTERNET PROTOCOL (IP) LOOKUP REQUEST

  • Publication number: 20160197878
  • Abstract: Various aspects of a method and a system to process a geographical internet protocol (IP) lookup request of an IP address are disclosed herein. The method is implemented in a first server. The first server receives a first request for the lookup of the IP address from a first computing device. In response to the received first request, data is retrieved from a cache memory. The retrieved data from the cache memory is communicated to the first computing device. The cache memory stores the data received in response to a prior received request for the lookup of the IP address.
  • Type: Application
  • Filed: January 7, 2015
  • Publication date: July 7, 2016
  • Inventor: MILES ZARN

SYSTEM AND METHOD FOR CONTENT DEPLOYMENT

  • Publication number: 20150339330
  • Abstract: Various aspects of a system and method for content deployment may include a server communicatively coupled to one or more client devices. The server may be associated with a web-based store. The server may be operable to receive information from a client device to request access to the web-based store. The server may be further operable to determine whether the client device is to be assigned to one of a plurality of channels of content access based on the received information. Each of the plurality of channels may be associated with a pre-specified version of content. The server may be further operable to randomly assign the client device to one of the plurality of channels based on the determination that the one or more client devices satisfy a set of parameters.
  • Type: Application
  • Filed: May 20, 2014
  • Publication date: November 26, 2015
  • Applicants: SONY CORPORATION, SONY NETWORK ENTERTAINMENT INTERNATIONAL LLC
  • Inventor: MILES ZARN