▎Augmentative and Alternative Communication(AAC) Dialogue System◦ Develop a system for AAC researchers to build dialogue device prototypes for research.◦ Implemented large (vision) language models to handle multi-modal data

▎AWS Billing Portal◦ Improved database performance to handle tens of millions of data per day◦ Assisted surveying AWS billing patterns to increase spread revenue by 10%▎AIOps ETL◦ Implemented automated ETL across 11 data sources for anomaly detection and alerts◦ Constructed visual report for analysis using Data Studio◦ Optimized data storage and querying with secondary indexing, slashing querying costs by 85%▎App Migration◦ Developed a GCP CI/CD service for automated code testing and deployment on Compute Engines▎Auto-reply Back-End Service◦ Replaced live customer service by building an auto-reply service with Flask API◦ Containerized and deployed the service using Docker to Cloud Run

▎Deep Relation Explorer Frontend◦ Developed a front-end for a research project which visualizes interactive network graph◦ Allowed user to input entities and interactively generate the results

▎E-commerce Recommendation System◦ Building a product label library with a pre-trained BERT model◦ Optimize the structure and code of the model, reducing the execution time by 10%◦ Create a pipeline from web scraping, data processing, and model training▎Advertising System Optimization◦ Upgraded to BERT model for article labeling, boosting accuracy from 50% to 88.4%◦ Refactored Spark-centric code to greatly optimize efficiency and increase speed by 13 times▎Post-buy Analysis◦ Analyzed invoice data to inform marketing on consumer habits and product relevance◦ Automated report generation process, saving 8 hours of manual work each week

◦ Taught English to students ranging from grades 3-6