DONG WEI LIN

AI / Data Engineering Projects

1. Product Categorization System Revamp

• Redefined the system architecture to replace the deprecated classification model
• Refactored the LLM workflow and improved overall modularity
• Rebuilt service logic and reorganized the codebase for better maintainability
• Optimized data pipelines by implementing advanced ETL processes and streamlining data flow.
• Deployed a Cloud Run service to support single-record inference
• Improved workflow efficiency and reduced cloud cost by USD 5–10 per month
• Enhanced system performance and increased processing speed by 22%

2. Product Image OCR Pipeline Revamp (1P / 3P goods)

• Migrated the OCR workflow to a new OS environment and rebuilt related components
• Implemented a scalable image crawling, downloading, and uploading mechanism
• Built a high-concurrency processing architecture to increase throughput
• Integrated GCP services, including Cloud Storage, Cloud Run, and BigQuery
• Designed a large-scale batch OCR solution using BigQuery + Gemini LLM + Object Table
• Optimized the data pipeline to significantly improve runtime performance
• Achieved 5× cost efficiency improvement and 10× faster processing

3. Misleading Advertisement Detection

• Collaborated with an external legal-tech vendor to validate compliance based on OCR text
• Built the data pipeline to send extracted text to the vendor and process returned results
• Designed internal reporting tables based on business requirements
• Revived a long-stalled project and drove it to the requirement-confirmation stage
• Improved vendor collaboration quality and upgraded to an advanced inspection tier
• Achieved an estimated 4× value increase based on contract cost comparison

Data Engineering

1. Built a Data ETL Pipeline with Python for the CRM project

• Customized the CRM data flow architecture and logic based on the Business Development (BD) team’s requirements
• Integrated the third-party platform HubSpot to gain more lead-generation opportunities

2. Optimized the reporting project by building a Data ETL Pipeline

• Optimized data processing by implementing efficient ETL pipelines and streamlining database design.
• Collaborated on ETL (Extract, Transform, Load) tasks, maintaining data integrity and verifying pipeline stability.
• Automated routine tasks using Python scripts, increasing team productivity and reducing manual errors.
• Used Airflow to establish a scheduling system
• Optimized code to reduce time complexity and increase execution efficiency
• Addressed technical debt issues
• Integrated GCP services to build a Data ETL flow connecting GA4 and BigQuery

DevOps

1. Infrastructure Setup and Maintenance

• Utilized AWS EC2 to set up a new environment
• Deployed necessary services on EC2 using Docker
• Implemented AWS Athena to handle semi-structured data

2. Established CI/CD Processes

• Used CodeCommit for Git version control
• Employed CodeBuild for continuous integration
• Adopted CodeDeploy for continuous deployment
• Built a CI/CD pipeline using CodePipeline

3. Cloud Service Deployment, Migration, and Optimization

• Deployed AWS services using CloudFormation templates for easier subsequent management
• Migrated some existing services to the newly built infrastructure
• Reduced service costs by approximately 25%

ESG Project

• Tree Identification in Satellite ImageryEmployed OpenCV image processing techniques and machine learning to identify tree canopy sizes
• Designed a calculation model to convert identified tree canopy areas into carbon neutrality metrics

Digital Transformation Project

1. Intelligent Textile Dyeing

• Consolidated past manual and digital data to create an Entity Relationship Diagram
• Planned operating SOPs for dyeing machines to reduce inconsistent operational errors
• Built an automated Data ETL process to minimize human error
• Developed a linear regression model for dye recipes to optimize formula adjustments
• Assessed and forecasted production costs

2. Assisted in Planning the Group’s Carbon Footprint Platform

• Designed the overall platform architecture
• Planned the data architecture and data pipeline

Lead a Team

1. Scrum Development

• Defined specifications and assigned tasks
• Iterated on project updates and conducted retrospectives for improvements

2. Led a Study Group to Explore New Technologies and Apply New Tools in Development

• Studied MongoDB technologies
• Researched and discussed Kubernetes
• Technical Support
• MongoDB Customer AssistanceProvided deployment assistance and troubleshooting

Ops

1. Maintained the Group’s Microservices and Project Code

• Deployed and managed applications on Kubernetes using Helm

2. Implemented Remote Backup

Data Engineering

1. Built an automated Data ETL and analysis workflow

• Implemented a SAS system to construct the game data ETL and reporting system
• Rebuilt and optimized the game data ETL using Python
• Provided data pipeline templates for subsequent projects

2. Led a Machine Learning project to establish a player tagging system

• Utilized machine learning algorithms from Python’s scikit-learn to assign tags to players
• Leveraged the player tagging system to help game designers and marketing teams set game parameters and plan marketing activities
• Developed a predictive spending model to forecast when players would top up, increasing the likelihood of in-game purchases
• According to operations management, integrating this system increased revenue by 15–20% over two years

3. Designed, maintained, and developed SQL and NoSQL databases

• Designed, maintained, and developed APIs related to MySQL (SQL)
• Planned the data architecture for Elasticsearch and MongoDB
• Designed, maintained, and developed APIs for Elasticsearch and MongoDB
• Designed, maintained, and developed Kibana dashboards
• Trained and assisted colleagues in using the ELK stack

Backend

• Assisted in the development of game systemsUsed Node.js to develop the game backend system (data layer)
• Employed Node.js and Python to develop the game ranking system

• Developed the image recognition system.
• Developed the speech recognition system with google API.
• Optimized the Chinese word segmentation system.
• Built the dictionary for word segmentation.
• Analyzed the word and recommended new word system
• Managed the E-commerce DMP (Data Management Platform)