Richard Jefferson Email and Phone Number

Engineering Productivity, often referred to as EngProd, is a Google-wide engineering discipline focused on accelerating Google innovation and development, maintaining engineering excellence, and enabling productivity at scale.EngProd teams partner with product development teams such as Search, Ads, YouTube, Android, Google Cloud, Shopping, Travel, and News. Each EngProd team specializes in creating systems and developer experiences that are fast and easy, resulting in quality products tailored to meet the needs of their partner team. EngProd roles primarily include Software Engineer (SWE), Test Engineer (TE), and Release Engineer (RE) -- although EngProd includes many other disciplines as well

Developed microservices that made up the business-critical ArchLink system used to track customer accounts, policy submissions, and policy producers; flag entries that may be duplicates; and perform regulatory checks, such as OFAC (Office of Foreign Assets Control) sanctions compliance. Migrated existing services from manually deployed Java JAR services on Azure VMs to containerized services deployed into AKS via a modern CI/CD pipeline composed of Jenkins and Harness.io.Increased the velocity of integration with newly acquired companies by developing a load-test suite using ReadyAPI to identify performance issues with existing microservices and then correct them with Java and SQL changes, reducing the API execution time from a 10-second average under load down to 5 seconds. Accelerated regression test process from 5+ days to 2 by assisting QA with creation of automated test suites and engaged with architecture and purchasing teams to begin building facilities for automated integration, smoke, and performance tests of services within CI/CD.Enabled part of the strategy to double Arch’s core insurance business by speeding the clearing process, which was bottlenecked by a slow 5-second SQL query, and replacing it with a sub-second Elasticsearch cloud-based solution. Ensured the new clearance search would meet future needs by engaging with ES professional services to verify we used best practices on cluster configurations and getting input on improving query logic and data schemas.

Lead a portion of our new Network Management product based on a microservice architecture with the responsibility of distilling requirements from marketing to a team of front-end and back-end developers. Tasked with determining additional use cases from internal customers and other services. Also participating in back-end development which is a Spring Boot application exposing a REST interface (documented via Swagger) which interacts with DBs (SQL, Mongo) and other services/hardware via their REST interfaces.Participated in re-architecture of the new Network Management product based on initial feedback on first design. Developed Streamer module with RPM and Docker install options for insertion onto our RNI systems to deliver telemetry from network modules and metadata from the RNI database to our microservice cluster via Kafka for ingestion. Created test cases as needed for our CI/CD pipeline and maintained Jenkins configuration for Streamer branch check-in validation and automatic artifact release. Implemented full stack features that spanned across Streamer, Message Ingestion, Processing, MongoDB, and UI.

Spent a year evolving the flow control portion of the traffic scheduling system to better account for message mix and greatly increase in bandwidth utilization. This required socializing changes with, and getting buy-in from, other RNI components and external entities involved in the transmit path and to coordinate during the development phase. Implemented these changes, completely re-working the message acceptance criteria to include a detailed simulation of queue processing to more accurately forecast actual transmission time which took into account more factors (some new) from downstream entities. The previous queue structure was replaced with a single queue using a dynamic priority value to automatically elevate messages approaching the end of their SLA. A token-based flow control mechanism was added between the RNI system and the downstream egress points to assure messages flowed as quickly as possible without overrunning the far-end buffers. These changes leveraged many REDIS features to provide the required performance while keeping operations atomic and coherent. Participated in a small team to evolve the RNI for future large-scale Domestic and International opportunities. Helped shape the architecture and explored ideas such as moving to a model of shardded processing engines and moving from ActiveMQ to Kafaka as the underlying messaging service. Developed our custom controller for allocating and balancing shard assignments for proof-of-concept demos.

Developed infrastructure for the Sensus RNI (Regional Network Interface) used to communicate with smart grid endpoints (water/gas/electric meters, smart lighting, industrial automation controls). Worked specifically with the Arqiva RNI for the United Kingdom's Smart Grid roll out which is made up of over a dozen VM nodes, each running multiple JVMs, and capable of supporting thousands of messages per second destined to over two million endpoints.Implemented portions of a basic monitoring service that observes various system processes and services and can alert administrators of issues and/or take independent remedial action. Worked to enhance installation written using a combination of Chef/Ruby, RPM creation, and bash scripting.Completely re-wrote traffic scheduling subsystem that was originally outsourced to provide proper traffic shaping and rate limiting through the system with additional feedback mechanisms to slow incoming traffic when necessary. Implemented a priority-based, weighted, round-robin queuing mechanism for messages. Due to the distributed nature of the system, queued messages were stored in a shared REDIS cache. Queue status information and queue operations were performed via a combination of atomic operations and Lua scripts. Subsequently enhanced this service to provide basic SLA validation as part of message acceptance criteria. Also adapted to interact with certain special message transfer situations such as fragmented messages and sessions downloading firmware updates to endpoints.

Replaced expensive third party ANSI ISUP protocol stack on BTS-10200 softswitch after Cisco's acquisition of IPCell. Allowed for the rapid internationalization of the BTS.Rewrote and restructured entire protocol code base to allow close to 90% code-reuse between the hundreds of country variants while providing cleaner code.Developed various memory and resource management features on the new line of multi-core ISR routers. Maintained and extended packaging and installation for all ISR routers.Designed, coordinated, and delivered multiple demos of progress of ISR work status to various internal audiences. Ported and restructured IOS features to work on the CSR-1000 virtual router for virtualized environments such as VMware.Coordinated with multiple remote teams in San Jose, CA, Australia, and the British Isles among other locations.

Developed various telephony protocols (ISUP, TUP, PRI, H.323) on the Protocol Gateway (PGW) which provided dial offload and circuit switched-to-VoIP gateway services. Used a proprietary fourth-generation language (MDL) designed for message processing and state machine handling. Took the initiative to create an automated regression framework to verify all unit tests against the code base on a nightly basis. Worked directly with customers/support engineers for many rapid deployments of new protocol variants.Coordinated with remote teams in Richardson, TX and England.

Responsible for the E911 and B911 software code bases and performed the Y2K compliance certification on them.Participated in the National Emergency Number Association (NENA) trade show for Public Safety/911 vendors, service providers, and operators by manning the company booth and giving information and demos to attendees. Developed an non-traditional AIN-based system for handling routing of cellular 911 calls. Worked with third party partners to specify, prototype, and demo a working system.

Developed, enhanced, and maintained suit of voice dialing features.Restructured and corrected interactions between switch based features and Advanced Intelligent Networks (AIN) 0.1 distributed feature environment.Participated in forums on Local Number Portability (LNP) and was took the lead for breaking down specification requirements into internal requirements and work items for the multiple teams addressing the implementation effort.

Developed test tools for DMS100/DMS10 telephony switches that controlled mix of heterogeneous boxes that generated traffic while presenting a consistent abstraction to the user.Developed portions of a feature that used a proprietary protocol which allowed telephony users to connect to an Intelligent Peripheral that supplied interactive voice recognition for configuring a user's service such as voice dialing, call forwarding, do not disturb, etc.Designed a feature to allow any user to access interactive voice services via a call-in number with the ability for the intelligent agent to re-direct their calls.Lead a team of engineers to implement this feature while still a co-op.