• Designed and maintained high-throughput analytical services built on gRPC, ClickHouse, AWS, and Kubernetes, supporting complex SQL generation for customer-facing analytics and reporting queries.

• Led query optimization, benchmarking multiple execution strategies and resource configs; improved critical analytics query 5.7x faster and reduced memory usage by 2x under the same concurrent load.

• Developed a query performance testing framework and production-like load testing infrastructure (Locust + synthetic workloads) to validate performance at scale before release.

• Delivered async service clients and concurrency improvements; achieved ~2x throughput gains.

• Implemented and productionized customer-facing analytics features: multi-touch attribution, conversion reporting, portfolio reports, and cohort reporting.

• Designed and implemented automated integration tests in CI pipelines to validate end-to-end service behavior (gRPC APIs, SQL generation, and query correctness), significantly reducing production regression risk and improving release confidence.

• Built Airflow DAG pipelines to automate data warehouse hygiene and lifecycle management.

• Led incident mitigation and reliability engineering, including large-scale data cleanup (300M+ rows) in distributed ClickHouse clusters, EBS volume upgrades, dashboards, alerting, and operational runbooks.

• Collaborated with machine learning engineers to integrate deep learning inference pipelines into web services with AWS, such as EKS, Lambda, S3, SQS, RDS and ElastiCache, making them production-ready.

• Led a project of building a web application as an internal tool that significantly improved data operation efficiency for the Data team by 3x. Responsible for designing the user interactive logic with two front-end engineers, setting a timeline and milestones, and developing the backend system of the application with Python Flask.

• Refactored the architecture of a core AI service by converting a monolith into microservices, improving scaling performance by 2x on Kubernetes.

• Designed and implemented a CI testing architecture to prevent blocking among multiple services’ end-to- end tests, increasing developer productivity.

• Designed and implemented backend machine learning web APIs for computer vision web services

with Python Django, MySQL, RabbitMQ, AWS ECS and S3.

• Refactored cloud deployment architecture, making AI services capable of handling the same throughput with lower cloud computing cost by over 10%.

• Implemented Gitlab CI/CD pipelines to do integration tests, Docker containerization and deployment of different environments on AWS ECS.

• Maintained and monitored service health through AWS CloudWatch metrics, logs and alarms.

Built data pipelines to perform ETL in Python, including large-scale data processing and loading into a cloud data warehouse with a self-designed schema for deep learning model training.