Paul Winward Email & Phone Number

Tempe, Az

Chandler, Az

-- As technical lead, I provide technical guidance and mentoring to the team, coordinate feature development across various teams and staff engineers, drive down technical debt, and lead discussions about architecture, features, defects, and on-call responsibilities. -- Co-designed and helped implement critical backend Google service for GDPR compliance. This inexpensive, low-latency service gets about 1.5-2 million requests per day. I also helped improve performance via bulk processing and analysis of various hardware and algorithm configurations.-- Co-designed and implemented a backend Google service for tracking user activities in the app (sending an email, creating a task, etc.) To reduce costs, I introduced batching of events, data structure changes, and thread management. The changes resulted in a 200x decrease in calls to the service with only a modest and acceptable additional latency. Later, when a schema change was necessary, I successfully transformed 300 million Google datastore entities into the new schema with a custom Dataflow script running over a Hadoop cluster.-- Designed, coded, and tested numerous Java APIs related to contact management, sales opportunities, GDPR compliance, partner tooling, and user activities.

Mesa, Az

-- Developed secure RESTful Web services in Java with Spring Boot to generate and execute optimized SQL queries based on user choices with a reporting tool for a customer handling hundreds of thousands of MySQL records. Utilized graph-based algorithms to determine shortest, reachable paths from one table to another based on customized join conditions. Tables, filters, joins, and selectable columns with built-in support for various aggregate operations could easily be added after installation. Built with Maven and deployed via Tomcat. -- Wrote efficient, multi-threaded, Java-based Web Services for flagship product to regularly import half a million GIS records into Oracle-based IBM Maximo using an XML API. Code and database enhancements helped realize rates ten times faster than previously seen. Deployed to WebLogic and distributed WebSphere application servers.-- Spearheaded efforts to adopt best practices for the application development life cycle in the areas of version control (Git), unit testing (JUnit, Mockito), documentation, and coding conventions.

Tempe, Az

-- Designed, implemented, and tested numerous Web application enhancements for Seller Central that spanned both the front-end UI work (javascript, jQuery, JSP, HTML, CSS) and the back-end database design and business logic (Java, Spring MVC, Hibernate). Frequently relied on various AWS products (EC2, RDS, S3, DynamoDB, Amazon ML, SWF) and internal data stores (ETL/data warehouse, caches) for storage and processing. Services were deployed over Tomcat and required low latency, high availability, horizontal scaling across multiple data centers, and automated monitoring and reporting.-- Helped design, implement, and scale critical data mining algorithms in Java for brand identification on big data and associated Web services. Improved algorithmic scalability of department’s critical brand identification module by an order of magnitude.-- Redesigned team’s core storage (using an internal cache for Amazon’s JSON-like, ION-formatted data) to address business logic deficiencies. Successfully led transition to new data schema in design, implementation, testing, and deployment -- which included dual-system operation before final adoption of new system.-- Integrated internal branding QC application with seller support Java APIs to enable customer support representatives to see captured communication between application reviewers, brand owners, and CSRs.

Scottsdale, Az

I was a senior software engineer for a software company that provided diagnostic and prognostic tools for aircraft engine health management. I designed and wrote various machine learning algorithms for isolating root-cause failures among a multitude of inter-connected engine components, each with its own set of failure symptoms. I created false alarm-reduction algorithms for critical system failures in FedEx engines. I also led development on an Intel home energy analytics project. The company was later purchased by Intel.-- Designed and developed failure isolation algorithms for Navy aircraft engines using adaptive naive-Bayes classifier variants-- Led analytics development team on Intel Home Energy Dashboard prototype and completed biweekly deliverables on time for partner companies-- Created iOS applications for engine sensory data alerts, failure injection simulation, and distributed engine controls as POC demonstrations, wowing customers with functionality and interface design -- Enabled false alarm reduction of FedEx engine no-dispatch faults via VADAR visual programming tools to slim maintenance costs and ground time

I wrote mission-critical code for multiple in-house software systems. I designed and built an automated pipeline to submit and track a two year backlog of data files, bundled with dynamically generated XML metadata, to a public repository. The pipeline operated over an SGE cluster and communicates with a normalized Oracle database I designed. I designed and developed C++ caching classes for a tool with millions of inputs. I also enhanced some DNA error-correction algorithms via the incorporation of conditional probabilities and helped streamline the complex process of short-read DNA assembly.-- Designed, developed, and tested storage and caching classes for next-generation genome assembler using the C++ Standard Library and Boost libraries-- Engineered a distributed automated system to track and submit a two-year backlog of DNA sequence to public repository-- Created Perl/SQLite analysis framework for assisting management in $1.2 million purchase-- Designed and implemented new schemas, PL/SQL packages, and Perl code base for major LIMS rewrite-- Overhauled freezer system code and Oracle schema to provide a single unified system that serves four different teams and handles 30K transactions/year

-- Extended complex optimization algorithms for trading company using Visual C++ .NET-- Improved usability, stability, and security of over 80 Gentoo, Debian, and Windows machines-- Published three peer-reviewed papers in major conferences-- Validated and expanded theories of genetic algorithm problem difficulty in thesis-- Led development and customization efforts on CMS-based website

-- Designed, developed, and tested extensive PHP/MySQL general-purpose survey tools-- Provided expertise and direction to development team-- Collaborated with project leaders to identify client needs-- Streamlined project development through effective methodology and communication

M.S, Computer Science

Activities and Societies: Research Assistant for Prof. David Goldberg researching genetic algorithms Thesis: Fluctuating Crosstalk, GA.

B.S., Computer Science