Li-Hua (Henry), Liu

    In a software R&D department, my responsibilities include developing and maintaining the web portals / Restful API for DNS Zone management / protection, DNSSEC Key management, and geolocation routing.

• 70% back-end (NodeJs, php, Spring Boot) 
• 30% front-end (VueJs)

Accomplishments:

• Designed and developed OpenAPI, enabling customers to perform operations provided by the UI through the API. 

      (Spring Boot / Php / NodeJS, Project leader)

• • Conducted technical discussions with team members
  • Proposed a batch update approach enabling multiple records to be updated in a single request, significantly reducing database I/O load.
  • Proposed appropriate search methods
    • Assessed complexity and I/O operations
    • Evaluated request size for each API call
    • Enhanced API response time by reducing it by an average of 1 second 
  • Refactored a complex UPDATE API with multiple responsibilities into smaller APIs
  • Proposed appropriate methods for parameter validation.  
  • Contributing
    • A successful 7-year DNS Protection contract, outcompeting other major industry vendors.
    • Assisted a global gaming platform in transitioning from a legacy DNS provider, delivering web-based Record/Zone management APIs and supporting the onboarding of 21,000+ protected DNS zones.

• Designing and Implementing scalable and high-Performance Kafka Consumer for Audit Log Query Processing 

      (Php / Kafka / MongoDB / Docker)

• • This project involved developing Kafka Consumer to query audit logs from MongoDB  and respond to the frontend.
  • Problem Identification & Resolution: Identified and resolved performance bottlenecks in Audit Log queries caused by varying data volumes and long processing times.
  • Parallel Processing Architecture: Engineered a multi-process architecture with independent Docker containers for Kafka Consumers, overcoming PHP 5.x multi-threading limitations to achieve parallel processing and resource isolation.
  • Partition Assignment: Assigned specific Kafka partition IDs to Consumers via environment variables and Shell Scripts.
  • Performance & Stability Improvements: Significantly improved Audit Log query processing efficiency, response speed, and system stability.

• Design and Lead Backend Development of Access Log Scheduled Processing System

     (Spring Boot, MongoDB, AWS S3, SFTP, Project leader)

• • Independent Project Takeover: Proactively took over a Spring Boot backend system from another team and delivered new features within one month, demonstrating strong learning agility and initiative.

  • Scheduled Task Pipeline: Designed a reliable backend process for generating and delivering audit logs to S3/SFTP with data integrity guarantees.

  • Data Flow Management: Managed daily OpenAPI polling and MongoDB checks for consistent data flow and real-time status updates.

  • Project Ownership: Owned the full development lifecycle from requirement gathering to system design and deployment.

• Spearheaded the upgrade of the DNS Protection web portal to support new interactive dashboards and configuration interfaces.

  (Node.js / Vue.js, Project leader)

  • contributed to a $9,000 USD. /month commercial contract by delivering interactive dashboards and configuration interfaces.

  • Defined API request / response structure and coordinated closely with Data and DNS System teams, including URL design, query strings, and JSON format converted from Protobuf.
    

  • Designed a new configuration schema and enhanced the commit procedure to deliver changes reliably to DNS infrastructure.
    

  • Communicated directly with Design to handle evolving UI requirements
    

  • Surveyed and implemented advanced Highcharts features such as scrollable datetime axes, dynamic boundary markers, and stacked/bar chart types.
    

  • Developed dashboard-related backend APIs to power visualizations based on real-time and historical DNS data.

• Zone Owner Verify Process Optimization & Pending Zone Mechanism

     (Php, Project leader)

• • Managed a dataset of 40,000+ zones, optimizing MySQL indexing and query strategies to maintain efficient pagination and filtering at scale, improve user experience, and ensure data integrity.
  • Led the redesign and implementation of the Zone Owner verification workflow to support the creation of "Pending Zones" prior to verification, eliminating redundant input and enhancing user flow.
  • Implemented mechanisms to ensure unverified Pending Zones were excluded from deployment, and automated the transition to Normal Zones upon successful verification.
• API Development and Security Mechanisms
  • Core API Development: Developed  API endpoints (Commit API, Authentication & Redirecting) to handle complex configuration changes and secure user access.
  • User Access and Permission Management: Integrated Redis for efficient permission validation and session management.
  • Secure Communication Mechanism: Implemented JWT to ensure the security of API communication.
• Configuration Delivery and Zookeeper Utility
  • Configuration Delivery Mechanism: Discussed specifications with configuration recipients and developed robust configuration delivery mechanisms leveraging Kafka and Zookeeper, ensuring reliable and consistent deployment of updates across various destinations.
  • Custom Zookeeper Tools: Developed Zookeeper utilities, including recursive node management (add/delete child nodes), and Upsert operations.

    4+ years in a software R&D department, I was responsible for:

• Developing, and testing the server-side core functionality of the Java-based NMS (e.g., Device Monitoring with SNMP & CLI), which supports monitoring OS and network device.

• Develop Java-based application to converting, parsing, sanitizing source data(ex: XML, log file) and send the outcome to destination(ex: web portal, Oracle DB), by using JDBC,  Process builder, JDOM,  and other libraries. 
• Collaborating with sales, OEM partner (ex: HP), and platform engineer.

Accomplishments:

• Improve the workflows of device status  monitoring  by using multi-thread scheduledExecutorService

  • Fix the bug of un-normal blocking caused by infinite Snmp walk request.
  • Fix the problem of blocking caused by delaying task which takes excessive time to complete.
    
  • Customized polling interval for each monitoring device(ex: 5 min for device A, 8 min for device B).
  • Maintaining thread status and handle the request of task reschedule/add/remove for each monitoring device.
• In a project of building CHT Security event monitoring system which earned over NT$ one million dollars, I was responsible for: 
  • Requirements analysis
  • Development of the application integrated with the HP Arcsight logger API:
    • Automatically download, parse XML data from the FTP server, which produces hundreds of event data per minute.
    • Reverse lookup by logger API for detailed information.
    • Sketch active path diagram using JGraphX.
    • Encode to base64 and upload to the web portal via HTTP protocol.
• In a project which earned over NT$ two hundred thousand dollars:
  • Design and develop a firewall logs filter to automatically backup, import log data with five hundred thousand records per file to Oracle DB by using:
    • Parse and filter the different types of logs by using util. regex
    • JDBC batch insert / PreparedStatement, commit to Db per ten thousand records. 
    • Call Oracle bulk copy command by using Process builder.  

• Designed new workflows of  the device performance monitor, which being used in the official version of NMS production:

  • Improve the device's discovery feature by adding Telnet connection checking into the existing SNMP event receiver.
  • Monitoring with different types of protocol, include Snmp/Ssh, VMware API, IcmpPing, DNS, and HTTP URL.

• Implement data pre-processes by Java, and compare the quality of review classification by a different classifier (ex: K-NN, Naive Bayes) using weka:
  • Using Porter Stemmer and NGram tokenizer of Lucene API to get tokens.
  • Filter tokens by performing DF feature selection, transform for TF-IDF vectors as the training set, train and calculate the precision and recall of classifiers by weka.