Xander Song

Xander Song is currently a Developer Advocate at Arize AI. Prior to that, they were a Machine Learning Engineer at test.ai from January 2020 to October 2022. While at test.ai, Xander launched a no-code tool for creating functional UI test cases ($700K ARR, seven enterprise clients and >700 public downloads) as a member of the core platform team. Xander also enabled robust navigation of app under test using an "application graph" with screen similarity algorithms to identify state and shortest-path and Q-learning to guide automation.

In addition, Xander led the engineering effort on a proof-of-concept to automate gameplay accessibility and load time testing for a mobile video game streaming platform. This resulted in acquiring the platform as a client. Xander also uncovered ~20 unique bugs for a video game streaming platform ahead of their public launch by testing at scale with Docker and Kubernetes, using >150 parallel runners to monitor ~100 games across ~20 data centers.

Xander introduced experiment tracking for screen element classifiers and built an internal data exploration tool for assessing regressions between classifier versions. Xander also increased throughput for batch experimentation 10x by building a scalable feature extraction pipeline using Apache Beam with Google Cloud Dataflow and leveraging containerized training. Finally, they automated and accelerated our core product's build process by 5x, using GitHub Actions to package our download client and build installers for Windows and MacOS.

Xander Song attended UC Santa Barbara, where they earned a Bachelor of Science in Mathematics. Xander then attended UC Berkeley, where they earned a Bachelor of Arts in Philosophy. Xander also has certification from Coursera in Deep Learning Specialization and Machine Learning.