Christopher Weaver Email & Phone Number

Lancaster, Pa, Us

Seattle, Wa, Us

San Francisco, Ca, Us

A technical lead of the group responsible for the correct provisioning of Twitter's backend fleet. Conducted research to make data driven decisions regarding capacity planning and developed solutions to improve tools used. Built collaborations across the organization in order to improve our product and increase impact. - Researched and identified variation causes in capacity results and implemented improvements to reduce them - Analyzed failover failures due to capacity - Technical lead of work group defining and deploying common graceful degradation constructs - Lead efforts to run stress tests in production for increased capacity planning accuracy - Discovered and working on remediation of issue reducing utilization by 10% across fleet

Philadelphia, Pa, Us

Integral leader on two teams developing several platforms all around improving customer experience. One of the teams is building multiple platforms that drive digital interactions with customers in real-time with measurments of efficacy built in with AB testing. The other team is building a platform that will maintain the current state of customers and another platform that is running models for millions of entities with @1B features per day. This is accomplished by consuming a massive amount of telemetry data realtime and running models against the ones that are considered to be in a degraded state. The platforms are using common technologies: AWS, Kafka, Kinesis, Flink, Nifi.

Philadelphia, Pa, Us

I am currently leading a team as a team lead developing a Comcast scaled streaming data platform. It has been deployed, using terraform, puppet and ansible, on AWS. The platform is a set of services built on Kafka using collector and aggregate sites with an http proxy fronting Kafka. The first iteration is scaled to handle 200 Billion messages a day, 4GB/sec peak and 4,000,000 msg/sec peak. The platform has been deployed and at peak is currently running 90+ Billion msg/day, @2,000,000 msg/sec and @1.2 GB/sec. I have strong team building skills focused on agile practices to build a collaborative cultureon the team.In addition I have worked on customizing Heka(Mozilla), using Golang, as a streaming data ingest point. I wrote several plugins to do Avro conversions on the edge nodes before forwarding the messages to Kafka. I also have designed simulations in Scala, Hadoop and Spark for the analytics team.

During my time at Dragonfly I have been primarily a Data Software Architect leading the design and development of the Data Swarm product, http://www.dragonflydatafactory.com/data-swarm. The primary tools and platforms I am using are Java, Storm, AWS, Mongo. Data Swarm will simplify buildng realtime processing on Apache Storm and Apache Spark. I am leading a diverse group of onshore and offshore developers in active development on this project. Acting as product manager I am also building the roadmap of technical features in the product. Acting as Scrum Master I am teaching the team the fundamentals of scrum as well as organizing the work in a backlog with sprints.

Denver, Co, Us

For the past two years I have been primarily working with big data technologies such as Kafka, Storm, R, Hadoop. I was tasked with creating a new way of reliably moving logs from hundreds of application servers to a centralized repository and Hadoop. I led the building of a new Log Delivery System with producers or entry points, a Kafka backbone and consumers or exit points. We built monitoring, alerting, self healing capabilities, configurability and extensibility into the different components. I am proud of leading the development and having been a part of building these components. During the Log Delivery System rewrite we were able to create several real time streaming analysis POCs. We were also able to repurpose the real time analysis POCs and use them as a elaborate monitoring and alerting system to detect host and stream failures. We also implemented a real time stream validator to ensure the messages pass certain standards. I designed a low latency data steamer that streams a subset of our internal generated data to several clients. The system was designed with high availability, dynamic flow rating, alerting, monitoring and self diagnosing components. The system was designed to provide configurable formats, delivery urls and delivery methods. There have been several Data as a Service initiatives this year. All of them utilized Hadoop to perform some heavy lifting ETL work and then providing some summary level analysis. The primary focus has been to utilize our search, route and other data to build services that will create income. Additionally my team is in the process of rewriting the ETL systems primarily using Python and Hadoop. We migrated all our jobs to Rundeck, a job manager. Several new ETL jobs have been added and I am writing them as java Hadoop jobs.

Denver, Co, Us

During this year I was part of a small team that started the first ruby/rails web applications at MapQuest, the local and discover apps. I had never used ruby before so I spent a week or two learning ruby paradigm. We began by writing the local and discover apps that involved bringing many MapQuest API’s together into a single page. We wrote many gems that wrapped the calls to these services. During this time I was able to work with some open source ruby contributors to improve their code. One of them was the http library typhoeus (https://rubygems.org/gems/typhoeus). I discovered a connection “leak” and was able to help fix it. On these projects and several other ruby web projects I worked with twitter bootstrap, rails and capistrano. With respect to testing we used capybara, rspec, selenium and jasmine. I was able to work with facebook auth, twitter auth and openauth for authentication on the discover website. As we wrapped up the local project we began to work on an oauth service for all our applications. I wrote a gem to be included in another project that was being written by an outside company. We were able to smoothly integrate the oauth gem directly with little effort on both sides. I was able to identify a product requirement that exposed an email harvesting risk during this time.I continued running SEO monitoring to determine errors and loss of rankings on all our website pages. I identified issues with Google, Bing and Yahoo web master tools. I was also responsible for working with offsite resources to migrate a set of legacy systems.

Denver, Co, Us

Reading, Pennsylvania, Us

Chicago, Illinois, Us

Chicago, Illinois, Us

Allentown, Pa, Us

Washington, Dc, Us