Andrew Wang Email & Phone Number

San Francisco, Ca, Us

San Francisco, California, Us

I support the Storage and Data Infrastructure teams at Airtable. We are responsible for providing powerful, reliable, and easy-to-use storage and query abstractions used by product development teams and internal data users. These abstractions are built on top of technologies like MySQL, DynamoDB, TiDB, Databricks, Spark, Kinesis, Flink, and Starrocks.As the manager, I am responsible for setting the vision and strategy for the teams, driving execution of key initiatives, and supporting and mentoring my reports. I work with cross-team and cross-functional stakeholders to align business and project priorities. I also run our engineering-wide incident management process and our engineering blog.Here are some recent projects my team and I have worked on:- An S3-backed database for transactional logging workloads, with 100x improved cost-efficiency compared to RDS MySQL- Migrating our analytics workloads from Redshift to Databricks- A real-time streaming and analytics platform based on Flink and Starrocks, which powers analytical product features for our highest-scale enterprise customers.

San Francisco, California, Us

I support the Storage team, whose mission is to accelerate product development at Airtable by providing scalable, reliable, and easy-to-use storage abstractions. Our primary storage system is MySQL, with Redis for in-memory caching and DynamoDB for key-value workloads.I grew the team from just myself to seven engineers. Here's a sampling of projects that we've worked on:- Upgrading from MySQL 5.6 to MySQL 8.0 (see: https://medium.com/airtable-eng/migrating-airtable-to-mysql-8-0-809f0398a493)- Launching a new key-value storage abstraction built on top of DynamoDB- Introducing TiDB for high-scale, business-critical relational storage workloads- Redesigning our ProxySQL infrastructure to 100x our MySQL connection scalability- Improving scalability of our read-heavy MySQL workloads with highly-available and low-lag read replicas

San Francisco, California, Us

I worked on the Data Infrastructure team, where I led the development of our feature flagging and experimentation system, developed and maintained data pipelines, and made improvements to our Airflow and cloud infrastructure.

San Francisco, California, Us

Scale provides high-quality training data for AI applications. I was the first senior hire in engineering, joining during the Series B.As a member of the Platform team, I was responsible for improving the infrastructure and code abstractions powering Scale's products, as well as our observability, build/test/release, and data warehousing infrastructure.Projects I worked on:- Monitoring and observability. I rolled out structured logging within the company and maintained our ELK cluster. Setup alerts and dashboards for our production systems and defined per-team SLIs and SLOs. Led migration of our monitoring from New Relic to Datadog.- Initiated use of Terraform for managing our cloud infrastructure. This allowed us to audit and version infrastructure changes and easily manage additional production and test environments.- Developed an automated daily release process and canary environment for testing with production traffic. This greatly reduced the number of site incidents, improved developer efficiency, and sped up testing feedback loops.- Led migration of our data warehouse from AWS Athena to Snowflake. I evaluated a number of different data warehousing technologies, coordinated the migration of our production data and query workload, and performed training sessions for dozens of SQL users at the company.

Santa Clara, California, Us

I joined Cloudera as approximately employee #300. This was an incredible experience at a growth-stage company. By the time I left, it was about 3000 people.HDFS (2012-2017):HDFS is the most successful open-source distributed storage system. I joined as one of three engineers developing HDFS at Cloudera, and became the tech lead for a team of ten.Some project highlights:- Led development of erasure coding, a new method of storing data that reduces storage costs by more than 50%. This was a multi-year project that involved novel research ideas and coordinating 20+ developers across multiple companies and multiple timezones. This is the biggest and most complex effort we've ever shipped in HDFS.- Release manager for Apache Hadoop 3.0, the biggest release of Hadoop ever. This was spurred by our desire to ship erasure coding, and required coordinating with multiple feature owners and downstream ecosystem dependencies over the course of more than a year of alpha and beta releases. I made substantial improvements to our release tools, process, and testing along the way. Hadoop 3.0 was the foundation of Cloudera 6, Cloudera's first major release in five years.- Became an Apache Hadoop committer, PMC member, and Apache Member due to my open-source contributions. Mentored multiple coworkers and community contributors into becoming committers and PMC members themselves.- Presented at conferences including Strata Hadoop World, Hadoop Summit, Apachecon, Apache Big Data, and USENIX ATC.Altus (2017-2019)Altus was Cloudera's Platform-as-a-Service offering, making it easy to run machine learning and data engineering workloads on the cloud.- Led development of a cloud-native machine learning platform built on Kubernetes.- Made improvements to Altus core services, observability, and build/test/deploy infrastructure.

Berkeley, Ca, Us

PhD track graduate student in distributed systems, advised by Professor Ion Stoica. Published papers at top-tier conferences (OSDI, SOCC) about in-memory caching for distributed filesystems and providing SLOs for heterogeneous storage workloads.

Master Of Science (Ms), Computer Science

Bachelor Of Science (Bs), Computer Science

Education record