Jake Kaufman Email & Phone Number

New York, Ny, Us

San Francisco, Ca, Us

Lead of the Developer Experience org. Responsible for keeping Lyft developers productive whether they are iterating on unit tests in their IDE or using the app to debug issues.

New York, Ny, Us

Team lead of large scale initiative to migrate from an aurora/mesos datacenter to kubernetes in AWS in order to meet growing needs of business while containing cost. Built, maintained, and sunset over the course of 5 years several distributed data stores aimed at supporting workloads not suited for mongodb. These “model services” are a system to export mongo’s replication logs to kafka and aggregate the updates in a stateful service to cache joins in redis. Built a read-only key-value store to deliver offline data to our online services called hfile service. Developed open-source replacement written in go called quiver. Rewriting in go lowered per process overhead enough to enable a single process per hfile. Wrote orchestration tool to automatically verify and deploy new data from our HDFS cluster. Built delta-service to combine offline data with mongodb. Quiver serves the majority of the data cheaply while mongodb stores write on a short TTL. Engineers use a reimplementation of our ORM rogue. Team lead on an experimental key-value store called “Kyiv” to replace our model services. Kyiv is a scala service that stored data in rocksdb. It also uses Kafka as a distributed write log, and aurora/mesos to provide scheduling. Gobblin was used to create rollups from kafka to allow kyiv instances to bootstrap. Rewrote image resizer service to lower machine costs. After 5 years of running in production, migrated to imgix to reduce overall cost and remove technical debt. Lead initiatives to raise standards around code review, launches, infrastructure changes, and post-mortems. Run a weekly meeting to facilitate cross-team discussion of large code or architectural changes. Member of our on-call rotation. Responsible for responding to alerts and resolving incidents. Provide direct mentorship to junior team members. Hold office hours to provide guidance on infrastructure to support new products.

Norwalk, Connecticut, Us

Contributed to application designed to calculate and analyze portfolio risk. Used boost libraries to extend large C++ code base deployed on OpenVMS and linux.Developed application to generate statistics based risk models using C++, json, and multiple kinds of databases. Work included creating a system for regression testing, and detecting updates to the underlying data.Developed a style guide for code and documentation that leveraged doxygen.Developed and taught a new bash lecture and lab for new hires.

Bala Cynwyd, Pennsylvania, Us

Developed a comprehensive database of servers and client interface to allow infrastructure teams to build dynamic server lists for remote maintenance jobs and monitoring; updated maintenance job launcher to work with new database system; result was to alleviate production failures due to outdated server lists and better tracking of failures.

Pittsburgh, Pa, Us

Developed labs and tests to expand and check student knowledge. Taught a recitation and 4 hours of 1-on-1 time per week to help students with homework. Students completed labs covering load balancing between remote servers in Java, a networked game, OpenMP and a Hadoop lab mining Wikipedia.

Pittsburgh, Pa, Us

Taught recitations and held office hours for an intro course to C and Unix skills.

Database administration and management to ensure the storage of important contacts; Website/form work to allow better interface with a Reaction contacts database.