Richard Kaye Email & Phone Number

Philadelphia, Pa, Us

New York, United States

When I arrived at HealthRhythms, they were struggling with application 'brittleness'. Even small changes to the backend resulted in a significant chance of breaking the customer experience due to a legacy architecture that lacked flexibility. Even migrating off of the old backend was hard, because they have customers who depend upon it, and expect it to continue working.I created a smoke test api suite in python (using pytest) that simulated mobile devices on their network, and exercised all of their api endpoints. Then I integrated the smoke tests with github actions, so that when they pushed a change the tests ran in the staging environment as part of CI. This tested the user experience before customers had a chance to, and found bugs in 'unexpected' places.I then extended the smoke test platform to be used as a stress test platform to help the team work through performance issues they were encountering with certain customers. My work let the team simulate very large batches of customers for the first time, identify bottlenecks, and formulate a plan of action to make the system more scalable.

When I arrived at Pie Insurance, I observed that there were challenges in how they managed Operational Data which were preventing the company from meeting it’s business goals and serving it’s customers. First and foremost, the pipeline for managing yearly rate updates from the state bureaus was almost completely manual. Each state update routinely took up to a week to process. The business requirement was to have all rate updates complete by the first of the year, but the company had never before completed all the updates on schedule. Failing to meet that goal created extra, unnecessary, work for our compliance department, which had to file for exceptions with the bureaus in each state in which we were behind. One reason the updates took so long is that the work had to be completed twice for each state, once for each backend that used the data.Second, the company needed to support multi-carrier to become profitable, and all of the regulatory and pricing data was stored in a database which had no concept of an 'insurance carrier.' The data was denormalized; adding carrier to the existing data store would have made updates even harder, and duplicated data all over the database. Leveraging my experience with automated testing, I lead a team of four engineers and together we built an automated pipeline. I was the 'tech lead' on the team; I coded the pipeline and managed the project.When we finished, the time to process a state bureau rate file fell from one per week to multiple states in a single day, most of which was testing. We completed the rate update cycle on time in 2021 for the first time! We re-modeled the data, normalized it, and added automated tests to prevent bad data from ever getting into the system. We made data updates so easy that our team was no longer needed; the resources on that team were moved to other teams.

Denver, Colorado, United States

Build, deploy and test system for managing Operational Data for the company.Created Checkbook application which enabled batch processing of policies into our system. This resulted in 12 million in new revenue the first year.

Edeliv TeamQA Automation EngineerScrum MasterIncident CommanderDeveloper on CallThe Email Delivery team owns AWeber’s high velocity email pipeline which sends millions of emails per day. When I arrived, they had an extensive unit test framework, but nothing that tested at the acceptance level or end to end. They relied on manual tests and getting push back from Email Service Providers to tell them if the mail we sent was well formed. Manual testing took a long time, because there were five different classes of mail which had to be verified. Bugs in this pipeline had an out sized impact on our customer because once a malformed email is sent it cannot be retrieved. While the traditional back end http API acceptance test pattern was really helpful here, the team also needed a way to verify emails after composition. The synchronous pattern of “run a test and validate a result” didn’t work here. Emails can take 10 minutes to arrive, and leaving a test open for that long consumes resources and leads to flappy results. To solve this problem, I designed and built the asynchronous test platform. A periodic process generated emails every few minutes. Another periodic process retrieved emails, parsed each email, and broke it up into its constituent MIME parts. These parts were encoded into an AMQP message, injected into our rabbitmq infrastructure, and then consumed as they arrived by my test consumer. The test consumer I wrote verified each message met the requirements in relevant RFC’s. Results were posted as metrics to graphite.The system tests every class of mail, email headers and message alternatives with the push of a button. We do this continuously and on demand. The system is effective at catching problems before they get to production. It has caught bugs which would have resulted in outages.

Horsham Pa

CP TeamQA Automation EngineerWhen I arrived at Aweber, they were struggling with a challenging transition to a new architecture. As part of the transformation of their product, they needed tests to make sure that as they made big changes that they didn't break existing customer facing features.This kind of problem can be mitigated by front end automation. We selected http://robotframework.org/. The test framework we built was an email validation system that sent 4500 emails a day, and validated reception. Test cases used BDD and employed gherkin. For this project, we went for breadth in stead of depth. Each test covered a key business flow through the system, and was designed to run continuously.An additional benefit of the system we built at Aweber, is that it brought to light bugs which had been missed during development and had hit the field. Bugs caused by race conditions, which are common in systems backed by asynchronous web services, were uncovered as the test framework continuously ran. Our framework caught pre-existing bugs that only happened once every 100 or 1000 customer interactions.

Malvern Pa

In 2014, Scala’s success in the small to medium enterprise market resulted in sales to much larger organizations. These new customers, including fortune 500 companies, were a potential windfall for Scala, but upon winning these accounts the the company discovered that we had challenges scaling the Content Manager RESTful API to meet the new demands of some of these larger deployments. In response to these challenges, I implemented a performance test framework based upon gattling.io. The framework enabled us to test the Content Manager API at a scale large enough to mimic the workflows of larger deployments. I worked directly with customers to understand their workflows, and wrote gatling scripts to mimic those workflows within the framework. The data I collected made it possible for our development team to target small changes to the Content Manager API which mitigated scaling issues for the customer until larger, architectural, changes could be implemented. As a result of my work, Scala kept its promises to these new customers, which will translate into hundreds of thousands of dollars of future revenue.Before my tenure at Scala, the test team for Content Manager consisted of manual testers only. My first task at Scala was to remedy this situation. I identified a toolset and wrote a framework in Python, using nose as the test runner, and requests as the http library. I wrote API tests in Python, and integrated them with a Jenkins Continuous Integration server for execution with each build. We went from zero to Auto in just a few months. Bugs caught by the new system were caught before the build entered manual testing, and before development moved on to the ‘next feature’. This drastically reduced turn around time for fixes to these bugs.

Princeton Junction Pa

When I started, the entire Triveni test process was manual, labor intensive and at times repetitive. I overcame this obstacle by creating a pilot test Automation program (Junit and Selenium). Creating a test automation program from scratch required several steps. First, I wrote the code for the test framework that provided the foundation for test automation, result reporting and logging. Second, I trained two of the three manual test team members in basic Java programming, so they could contribute to test suites. Finally, I configured a Jenkins environment to execute the suites periodically so we would never lose sight of the quality level of our software. The program was so successful that management decided to expand it.While at Triveni, I revamped the entire manual test process to bring test cases into alignment with product definition. This strategic change significantly improved final product quality, as testing concentrated on the areas of greatest customer concern and greatest development risk. The new process also moved test lab build out earlier in the development life cycle, thereby eliminating a common source of product delay.

Horsham Pa

Iron Core is a large test automation framework intended to be used by many diverse product groups within Motorola. The project supplied a common set of tools and API's to development and test teams throughout the company so that they could quickly implement test automation with common result reporting, analytics and resource management. The common platform saves time and money by eliminating redundant work; each project group leverages the toolset offered by the platform rather than re-inventing the wheel. As System engineer, I collaborated with development and test teams to identify product requirements for Iron Core automated test framework and wrote the Product Functional Specification. I designed automated test cases using AutoIT to support initial roll out of Iron Core framework. I set up servers for the framework and interfaced them with Motorola IT.Wrote functional requirements for NC2000 next generation IP to Aloha bridge. The NC2000 is fully redundant, which drives down operating costs, and it implements new licensing technology which drives revenue. Designed and prototyped network to support NC2000 using CISCO and Broadcom switches. Fully redundant network supports OSPF, PIM, and VLANs.Designed Systems engineering hardware/software lab from scratch. The new lab integrated OOB Aloha technologies with modern DOCSIS enabled systems. This lab provides a stable test platform for the systems engineering group to prototype new features.Maintained SNMP MIB for APEX 3000 EQAM. Oversaw the definition of the SNMP interface, and worked with developers to incorporate the MIB into an existing software design.

Horsham Pa

Designed and built Custom Transport Stream Generation station, which used 'off the shelf' amplifiers, transport stream multiplexers, and satellite receivers to quickly and easily assemble test streams with unusual MPEG characteristics. Reduced stream generation time from an average of two weeks to an average of three days. Received Motorola innovation award 12/09 for the Custom Transport Stream Generator. Created custom MPEG transport stream files to support the system test lab. These streams duplicated conditions found at customer sites, and enabled the test team to validate new requirements more quickly and with greater precision. End test lead for Digital Transport Adapter (DTA) project. Customized MPEG analysis tool to parse SCTE 65 tables. Reduced test cycle time by 25%. Test lead for Set Top Diagnostic Controller (STDC). Led team including three engineers.Tester on APEX 1000 release 2.3.8 and 2.4. Tested Apex in Switched Digital environment utilizing Motorola SVM and ERM.

Solicited customer network topology diagrams, analyzed their MPEG monitoring needs, and developed plans to meet those needs using Triveni products.Developed MPEG-2 training programs for Triveni Digital and the Society of Cable and Telecommunication Engineers (SCTE). Presented at SCTE Central Florida chapter in 2007 and for SCTE Cable Expo in Orlando FL, June 2007. Contributed to Standards development in the Advanced Television Systems committee (ATSC). Member of TSG1 for the development of A/78, the ATSC Recommended Practice for Transport Stream Monitoring.Increased sales by educating re sellers and performing demonstrations for strategic decision makers.Implemented CRM on Demand by Siebel for Triveni Sales force, provided technical support, user training, and sales process development for Triveni's first Customer Relationship Management system.

Ensured client satisfaction and salvaged a million dollar contract by working in a high-pressure environment on-site with the client to resolve a potentially disastrous situation with a BIU surge suppression circuit. Increased reliability and reduced calls to product support by developing a new BIU test plan, and verifying a new version of third-party firmware for CSU/DSU card.

Education record

Bs, Electrical Engineering

Activities and Societies: Eta Kappa Nu, Tau Beta Pi