About me

My name is Chengkai Yang.

I am currently a Master of Applied Data Science student at the University of Victoria, my cumulative GPA is 8.11/9

I have 5+ years of work experience in big data development and back-end development in the fields of autonomous driving, finance, and banking.

Employment History

Aug. 2025 – Dec. 2025 Victoria, BC

Designed and developed performance monitoring components for real-time streaming applications, collecting runtime metrics such as backpressure and throughput. Validated through testing to accurately identify operator-level bottlenecks in streaming jobs.
Built 20+ Grafana dashboards to improve observability for 100+ streaming jobs and help developers identify performance bottlenecks in production environments.
Led the integration of performance testing into GitHub CI/CD pipelines to enable fully automated performance testing. Worked with DevOps engineers to configure GitHub self-hosted runners and ensure reliable execution inside Kubernetes pods.
Implemented a GitHub bot to trigger performance tests and publish test reports directly to pull requests, enabling data-driven performance evaluation and faster feedback for developers.

Apr 2023 - Dec 2024 Beijing,China

Designed and maintained storage and query pipelines for trajectory points and images from thousands of autonomous vehicles. Leveraged Flink for real-time processing and Spark for offline analytics, with trajectory data stored in HBase, and used Python and MinIO for trajectory image processing and storage.
Built Spring-based backend services and worked closely with algorithm engineers to operationalize point cloud and high-definition map generation algorithms, resulting in over a 3× increase in map generation efficiency.

For more details about my employment history, please go to my LinkedIn page.

I am also a contributor to the community of the well-known real-time processing framework Apache Flink.

My contributions mainly include translating the official English documentation of Flink into Chinese, which helps thousands of Chinese developers use Flink better. Here is a main list of my contributions.

Translate “Metric Reporters” page of “Deployment” This article mainly introduces how to use monitoring tools such as Prometheus and JMX to better monitor the running status of Flink.
Translate “Extension Points” and “Full Stack Example” in “User-defined Sources & Sinks” page:This article mainly introduces how users can customize the Table API to read and write data to external storage in addition to the Table connectors provided by Flink officially (such as Kafka connector, JDBC connector).

Gpt-style SpamEmailClassifier, click here for more details.
Here is an short video of a Flink project I did previously using Flink to process log to find potential bugs in real-time

I use Java, Scala and Python, as well as the following technology stack.

I really enjoy doing weight training at the gym, and I train at the gym of the University of Victoria.