New York, New York, United States
Experienced software engineer with extensive experience in building distributed systems, concurrent systems, fault tolerant systems, web applications, databases, and applications in the cloud.
Working on distributed systems features in the highly concurrent C++ codebase of the core MongoDB server. Particular focus areas include sharding, fault tolerance, cluster management, and distributed algorithm implementation.
Research focused on high-performance distributed systems, distributed databases, consensus protocols, and concurrency control. Contributed to a custom research platform to study distributed database transactions. Worked on a concurrency control algorithm that uses distributed dependency tracking to order commits. Worked on the prototype of a distributed database built on top of infiniband providing low latency and high throughput data access and retrieval. Modified the etcd distributed key/value store to implement optimizations of the raft algorithm.
Demonstrated how to collect data from existing distributed systems to generate a reinforcement learning model and accurately estimate the policy’s performance without the risk or cost of deploying it on live traffic. Modified the Redis key/value store to incorporate our trained model to make cache eviction decisions. This work was published in the HotNets 2017 conference.
Worked as part of the platform services team where I have played a critical role in increasing the fault tolerance and availability of several key backend systems which support Blurb’s global retail business.
Provided custom mobile and web application development services Cloud automation and configuration. Custom software development.