Tim Martin Email & Phone Number

Redwood City, California, Us

Seattle, Washington, Us

Seattle, Wa, Us

I was tasked with replacing legacy search and analytics systems including researching new tools, providing business use cases and provide implementation overview.- Demonstrated distributed stream processing through illustrations, containerized services runnable on localhost and organized hands-on meetings to allow team to quickly adopt.- Designed analytics system based on Apache Kafka / Apache Storm to enable real time metrics and allow replayability in case of data corruption.- Implementation of new search API and background state computation / synchronization script. Used ElasticSearch as data store to reduce code complexity and improve reliability.

Seattle, Us

I worked on refactoring / replacing original analytics reporting API.- Refactored PHP reporting layer to use a stack of transformations which would process all business logic one record at a time instead of its original batch behavior which allowed more predictable memory usage.- Added instrumentation and monitoring to existing systems to allow quicker response to system malfunctions and overload.

Seattle, Us

I was one of the original backend engineers on the MobileAppTracking mobile app analytics platform.- Wrote original CRUD API for using combination of open source tools and custom built ORM.- Wrote scheduled tasks for background analytics reporting jobs.- Leveraged AWS tools (S3, EC3, DynamoDB) to increase application scale with reduced operational

Seattle, Wa, Us

Lead development role continuing existing symfony project where my biggest contribution was optimizing MySQL schemas. After load testing with siege, I increased performance and reduced table size 10x on slowest part by denormalizing several one to many relationships in MySQL.Also, in a separate DB in a related project made dramatic improvements in write speed to biggest table in DB by dropping indexes and making the only field ever searched part of the primary key. Table shrank from 12GB to around ~6GB (table had around 300m rows), write speed increased at least 20 fold. Average row length from ‘show table status’ was cut in half. Implemented development tools for a more team based environment including connecting project svn repository to trac. Ran exploratory load tests with siege and performance profiling with xdebug to find the slowest parts of the framework. Deployed infrastructure that met project requirements in Rackspace cloud. Added instances of debian linux for 1 load balancer (pen), 2 web servers (apache, php), 1 master DB server (mysql), 1 slave DB server (mysql) and a project / development server (svn, trac). Tools used: PHP5, MySQL5, Apache2, Debian linux, xdebug, kcachegrind, Memcached, symfony, jQuery / JS, HTML, CSS, subversion, trac, siege, firefox plugins (web developer, HTTP Fox, firebug), IE, Safari, VMWare, pen, ssh tunnels, nagios, shell scripting, scp, ssh, vim, ZDE

Eden Prairie, Minnesota, Us

Application was very unique in that it was hierarchal when displaying data meaning when a user was viewing data they would see their own data aggregated with all child customers / users data (unless they restricted their view). Alertlogic analysts would see data for all customers /users since they were at the ‘top’ of the tree.Co-wrote UI search layer written in PHP which turned all search (which is the key part of the product) parameters into objects with recallable histories, search filters, interactive date ranges and the ability to be saved as a ‘view’ that could be re-used / extended later or shared with others.I created an interactive flash graph library (graphs rendered solely in actionscript) that allowed users to drag sliders to display subsets of graph data in greater detail. Any action on graph would communicate to a JS listener which would send a request to web server via AJAX using prototype library, get JSON response and update the page content corresponding to the graphs controls. Also created an export wrapper around FPDF and CSV to convert UI report / view output into PDF format. Output routines reduced to a few simple calls to create PDF and CSV output. Tools used: PHP5 (apache_mod and CLI), MySQL5 (including EXPLAINS, slow query optimization) / MySQL Workbench, Apache2, xdebug, kcachegrind, Debian linux, Memcached, Flash, Propotype / JS, HTML, CSS, Phpunit, Phpcs, Phpdoc, cruisecontrol, PEAR (HTTPRequest, Mail), subversion, trac, selenium, gearman, siege, firefox plugins (web developer, HTTP Fox, firebug), IE, Safari, rest / JSON web services, minify, scp, ssh, vim, ZDE, Photoshop, gd, fpdf, jpgraph