Logistic team

●Reinvent Project - Image SigningCollaborated in a team to design and implement signature deletion-related services, including the customer-facing API, Lifecycle policies, an asynchronous process for database cleanup, and database schema and table design.Enhanced service performance by applying parallel programming & batch processing, achieving a 41% reduction in service latency and supporting over 500K monthly transactions across 23 regions.●Tech Lead - CreateImage API v2Led the migration of database systems from JournalDB to DynamoDB, spearheading improvements in API designs and data structures.Dramatically reduced latency by 79% for CreateImage API by replacing table scanning with more efficient NOSQL queries and the best practice of single table design in dynamodb●Solo Project - Disaster Recovery PlatformDeveloped a platform providing an internal-facing API for atomic restoration of customer data across S3, JournalDB, and DynamoDB.Successfully restored over 1.4K repositories and 100K images for critical clients like Airbnb and Salesforce, enhancing data reliability and service trust.●Solo Project - Reaper's Horizontal ScalingEngineered a sharding algorithm that significantly increased throughput by approximately 1500%.Reduced data retention times from 7 days to 24 hours in major regions, optimizing system efficiency and resource use.●Solo Project - Analytics Service TestsDesigned and developed comprehensive testing protocols for a metering service, including sanity, integration, and end-to-end (EtoE) tests.Implemented automated alarms and rollback mechanisms to address deployment issues.Eliminated Severity 2 incidents and mitigated customer impact from future deployment issues.

● Brief: Solved an online problem in distributed, multi-tiered Alibaba cloud workflow system by implementing message queue utilizing MySQL and Tair(Redis) databases. This new-designed MQ model also greatly improve user experience in the other 13 APIs of Alibaba Cloud ECS Products● Process： Located the reason and searched for solution. Built class and data models and traded off the way to implement from time-consuming, success rate, and throughput. ● Achievement： Solution development: Running tests in simulated environments. Enlarged throughput from 0.14/s to 0.71/s when QPS = 50 served under 100 Threads. Reduced time consuming to a basically fixed value (± < 0.1s) and Increased success rate to 100%.● Extensions: Apply the same model to solve a superior problem. This model was also adopted by three P9 engineers to improve user experience over 33 resources and 223 APIs in the ECS product in the future.

● Brief: Replaced Custom Company’s ERP system with SAP S/4HANA Cloud and connected with others.● Process: Deployed the new system and trained key users. Adjusted the basic structure of the SAP according to users’ requirements and the features of the Chip manufacturing Industry.● Achievement: Implemented an independent material information collection website with Spring and Oracle.● Provided solutions in special services; involved in discussions and contributed possible plans.

● Analyzed multiple groups of condensate pipe rupture data and predicted timing for the next rupture.● Read and presented foreign machine learning journals and articles for other team members.● Reduced dimension by different machine learning algorithms. Tested each performance until found the most pragmatic one.