Li Yang(Jesse)

♦ Overview: Developed a comprehensive platform to manage relational database metadata, primarily focusing on tracking the lifecycle (creation, modification, deletion) of table fields in real-time. The system analyzes these fields to identify their business context (e.g., financial data, PII) and assess potential data risks.

♦ Key Features & Implementation:

• Real-time Tracking: Monitors the lifecycle of table fields, ensuring that any changes are captured and logged immediately.

• Automated Field Analysis & Risk Assessment: Implemented an offline data synchronization pipeline feeding an analysis engine. This engine employed statistical methods to determine field business types (e.g., order amount, fees, sensitive customer information) and estimate associated risk levels.

• Proactive Alerting & Reporting: Integrated an IM bot to deliver analysis summaries and risk notifications directly to subscribed users.

• Data Visualization: Built a BI dashboard leveraging the synchronized offline data to provide actionable insights into metadata trends, field usage patterns, and overall risk profiles.

♦ My Responsibilities: Independently completed the platform's backend architecture design, development, deployment, and maintenance.

♦ Impact & Usage: Successfully adopted by the entire financial business line with several hundred users, significantly enhancing data governance and risk management capabilities.

♦ Overview: Developed a robust system enabling users to perform near real-time data reconciliation and audits across diverse data sources. The platform supports checks for critical data points like financial transactions, order statuses, and other key business metrics, offering a user-friendly interface for defining custom audit tasks via simple SQL queries.

♦ Key Features & Implementation:

• Non-Intrusive Data Synchronization: Implemented a data pipeline using Flink and binlog-based ETL to synchronize online data into an offline HSAP (Hybrid Serving/Analytical Processing) data store. This approach ensured zero impact on the performance of production databases.

• Dynamic & Scheduled Reconciliation: Leveraged a scheduling platform to allow users to dynamically create and schedule reconciliation tasks based on their specific periodic requirements.

• User-Empowered Task Creation: Provided an intuitive UI enabling non-technical users to configure and initiate reconciliation tasks using straightforward SQL statements.

• Comprehensive System Monitoring: Integrated monitoring and alerting functionalities to ensure system health, data integrity, and timely notification of discrepancies.

♦ My Responsibilities: Design and implementation of the core data synchronization pipeline, ensuring efficient and reliable data flow from online sources (via binlog ETL with Flink) to the offline HSAP environment.

♦ Impact & Usage: Widely adopted by financial and e-commerce business units, significantly improving data accuracy and audit capabilities. And successfully processed and managed thousands of cumulative reconciliation tasks.

♦ Overview: Developed and implemented a comprehensive platform designed to mitigate the risk of online accidents caused by changes and updates across various internal company platforms (e.g., compute, dynamic configuration, gateway & router). The platform integrates into existing publishing processes to enforce controlled change behaviors and perform automated post-release checks, significantly enhancing system stability and reliability.

♦ Key Features & Implementation:

• Platform Integration: Seamlessly embedded change control and after-publish check steps within the UI and backend layers of diverse internal platforms (compute, dynamic config, gateway & router, etc.).

• Platform-Specific Checks: Provided tailored checking capabilities for different platforms, such as analyzing new vs. stale route differences for gateway & router platforms.

• Unified Checking Capabilities: Offered universal checks applicable across platforms, including automated test execution on changed servers and enforced canary publishing before full deployments.

• Unified Web UI: Developed a centralized web interface enabling business R&D and QA teams to configure control and check levels/strategies for their specific services and platforms.

♦ My Responsibilities: Served as a core developer, primarily focusing on the dynamic config platform and the gateway & router platform. Managed business connections, provided assistance, and handled on-call duties related to the platform.

♦ Impact & Usage: Successfully adopted by multiple major business lines, including Novel, E-commerce, and Toutiao News. Controls and monitors hundreds of individual service and platform instances, directly contributing to reduced online accidents and improved system stability across critical company infrastructure.

♦ Overview: Developed a comprehensive test framework and a suite of backend test tools to empower QA engineers with efficient test case management, flexible execution capabilities, and convenient access to testing utilities. This initiative aimed to streamline the testing process and improve the overall quality assurance workflow within the group.

♦ Key Features & Implementation:

• Test Framework: Implemented a mechanism allowing QA engineers to define and modify test cases using a single JSON format adhering to a specific JSON schema convention. Integrated with various environments, including CI/CD pipelines, for automated testing.

• Test Toolset Backend: Provided a batch of useful interfaces for testing purposes, such as retrieving online insight data and triggering data manipulation or recovery processes. Integrated with an Instant Messaging (IM) bot for flexible triggering and invocation of toolset functions.

♦ My Responsibilities: Accomplish the design, technical stack selection, development, and maintenance of both the test framework and the test toolset backend as a sole developer.

♦ Overview: Creating and maintaining a comprehensive offline test environment for our group, serving both Research and Development (R&D), Quality Assurance (QA), and external dependent business units. The environment simulated a complex production environment, encompassing nearly 10 microservices (both HTTP and RPC), intricate network topologies, and a wide array of dependencies.

♦ Overview: Building and maintaining Docker images for the stable offline test environment used by the Payment Department. The goal was to create a reliable and consistent testing environment that mirrored the online production system.