rafer@ralua.com
303.210.5970
Denver

Rafer Hazen

Overview

Experience building backend data systems with an emphasis on asynchronous messaging/queueing
Building, growing and managing engineering teams

Started the Data Pipelines team and grew it to 8 engineers
Lead the release of project Hydro, a Kafka-backed event logging, storage and processing system now used by 30+ teams
Hydro facilitates both asynchronous app-to-app communication and data warehouse ingestion and accepts ~250k writes/second at peak
Wrote Hydro's data warehouse ingestion component (event ingestion and schema management)
Lead the replacement of GitHub's background job system with a service that provides queue semantics over Kafka's log structured model
Wrote the job queueing system's sharding algorithm to distribute messages across an in-memory dispatch tier

Operational support and development of GitHub's in house-event data pipeline and warehouse
Evaluation, execution and support of Presto rollout
Developed application to acquire data for machine learning jobs and serve results to user-facing applications
Lead support and development of our internal A/B testing framework

Managed the API Team responsible for development, delivery and support of our group's public APIs
Implemented core data models and interchange formats
Delivered a performance-oriented overhaul of the infrastructure for our semantic categorization engine
Lead development of "Semantic API" to expose text-processing capabilities
Developed and delivered standardized "slug" deployment model to reduce errors and increase predictability
Rewrote legacy data ingestion framework to allow for horizontal scaling via the introduction of a queueing layer

Worked with project managers, developers and clients to complete projects
Spearheaded efforts to create and improve Foraker's culture of test-driven development
Developed interactive browser-based choropleth maps
Wrote an extensible framework to import, relationalize and map health insurance usage data

Maintained and augmented a Ruby ETL utility to warehouse search usage data from Apache log files
Assisted in research and development of Lijit’s authority ranking metric
Day-to-day maintenance and feature implementation for lijit.com

Built and deployed Oracle-backed PHP front end for an internal sales metric portal