Theseus is a distributed SQL engine that executes on NVIDIA GPUs. I• Led development of cluster elasticity to allow users to add new nodes to a running cluster.• Analyzed Total Cost of Ownership data with planner and execution teams to design performance improvement features for both cloud and on-prem deployments.• Led team developing SQL sessions, allowing users to move their working context between clusters.

During my service as chair, Impala democratically elected twelve community members to commit rights to the project, including the first two committers from outside Cloudera, and added three people to our governing body, the PMC. We have also released two minor versions and one major version, Impala 3.0.

Manager of the Presto platform team. During my tenure, the team built:• Integration to support geographically distributed tables within the data warehouse.• Support in Presto for new hardware and heterogeneous hardware.• Write coalescing that will save enough electricity to power half of a data center indefinitely.I also co-founded Data Infrastructure Diversity and Inclusion, engaging with groups like Girls who Code and minority-serving institutions like SJSU and Georgia State.

My team at Netflix is responsible for storage and metadata services around our data warehouse, which is a few hundred petabytes of Parquet files on S3. Additionally, we are building a one-stop shop for all business-critical metadata at Netflix, data both within and outside of our S3 data warehouse.

I co-founded and managed the Altus Data Warehouse team. I also managed the Altus Data Engineering team. These are cloud-based platforms for interactive queries and batch computation, respectively.

Apache Impala is a distributed analytical database management system for Hadoop clusters and cloud storage on AWS or Azure. Cloudera donated the code to the Apache Software Foundation, which is also the home of other big data projects like CouchDB, Kafka, Cassandra, and Spark.I led a team of 15 Cloudera developers through the Apache Incubation program for 18 months. This culminated in the project becoming an Apache Top Level Project. I am currently the Apache PMC Chair of Impala.I managed one Cloudera Impala release and two Apache Impala releases, during which the number of Cloudera customers using Impala increased by 30%.I built the keystone technology in Impala's runtime filtering, which helped increase Impala's performance on the TPC-DS benchmark suite by 300%.

Led Headrig, a near-real-time processing pipeline with thousands of distributed nodes as the team grew from two to five engineers. Decreased 99th percentile tail latency by 90%, incident rate by 50%, and false alarms by 90%.Led the design and construction of personal logs deletion-by-encryption, a storage and processing system for encrypting, storing, and processing petabytes of server log data, synchronized across thousands of miles.

Designed, built, and deployed the maps backend for Green Driver, a vehicle routing app. This work included building and deploying a multithreaded tile rendering and caching server.Spearheaded hiring efforts and biweekly tech talks.