About

Master of Computer Science @ Rice University Aug 2025 – Dec 2026

Master of Science in Computer Science @ National Yang Ming Chiao Tung University Sep 2024 – Jun 2025

Bachelor of Science in Computer Science and Information Engineering @ National Cheng Kung University Sep 2021 – Jun 2024

Software Engineer Intern @ Rivian and Volkswagen Group Technologies May 2025 – Present

Software Engineer Intern @ MaiAgent (formerly Playma) Feb 2025 – Jun 2025

Developed new RESTful APIs and features with Django while streamlining DevOps automation for the MaiAgent AI platform, gaining 6 new partners, including CTBC Bank, MSI, and HPE
Reduced CNAME vanity domain setup time by over 80% for our the MaiAgent AI platform by developing a backend service that automates CDN provisioning and SSL/TLS certificate management on AWS CloudFront
Cut production deployment time for 12 AWS EC2 instances from 30 to under 5 minutes by utilizing Ansible
Enabled air-gapped on-premises deployments for 3 banks by automating setup with Ansible, reducing deployment time from hours to minutes across diverse OS (Ubuntu/RHEL/CentOS) and VM topologies
Accelerated Ansible testing by 50% across inventories by building a CI pipeline with Molecule and GitHub Actions
Integrated HeyGen and D-ID services into the webchat to create talking avatars, presented to 10+ clients

Software Engineer Intern @ Trend Micro Jul 2023 – Aug 2023

Worked on the Web Reputation Service team, contributing to the migration of the Web Classify Server (WCS), a legacy service handling 700 million daily classification requests
Cut infrastructure costs by 10% by migrating the WCS legacy service from VMs to Kubernetes
Reduced WCS deployment time by 20% by building CI/CD pipelines with GitHub Actions and Helm charts
Improved service reliability by implementing a Prometheus exporter that monitored WCS pattern file updates across 200+ nodes, enabling real-time detection of outdated nodes and reducing reliance on user reports
Reduced desynchronization errors from ∼3 per week to nearly zero by redesigning the WCS pattern file sync process to address issues with AWS S3 Cross-Region Replication and network instability

Software Engineer Intern @ AInimal Jan 2022 – Jun 2023

Led the migration of a monolithic backend into 4 Go and Node.js microservices, designing gRPC APIs for inter-service communication to build a more fault-tolerant and scalable system
Deployed backend infrastructure on GKE with Cloud SQL and Cloud Storage using Terraform, improving scalability and reducing overall service downtime by 12%
Scaled chat service to handle 5,000+ concurrent users by adopting Go, MongoDB and Google Cloud Pub/Sub
Improved notification responsiveness by 20% by developing the notification service in Go and redesigning the click-through flow in React Native
Built test suites with 90%+ coverage and automated testing/build/deployment pipelines using GitLab CI/CD, reducing regressions by 16% and release time from hours to minutes

Software Engineer Intern @ Intelligent Mobile Service Lab Jun 2022 – Jan 2023

Refactored and optimized an online judge website with Vue/Vuetify, improving responsiveness and usability for 200+ students
Optimized core components with asynchronous loading, reducing average page load times by ~20%

Research Assistant @ Software Engineering and Intelligent Test Automation Lab Oct 2021 – Oct 2024

Developed a prompt optimization method for large language models (LLMs) using pairwise feedback, reducing annotation effort by 50% while maintaining performance
Implemented a new screen recording feature for Rapi (web application testing software), enabling local and remote browser recording during test cases which reduced debugging and bug reproduction time by 50%
Built an auto-scalable tool for a Kubernetes-based Selenium Grid, enabling reliable video capture across large-scale parallel test suites

Remote Research Assistant @ XR for Intelligent Medicine Lab Dec 2023 – Sep 2024

Developed a customizable LLM-based virtual therapist using LangChain, retrieval-augmented generation (RAG), and prompt engineering, improving personalized mental health support
Built a pipeline to preprocess and embed data from 3 mental health datasets of varying sizes (∼50k docs), storing embeddings in Qdrant for RAG context retrieval
Surveyed and experimented with different evaluation methods to assess the effectiveness of an LLM-based virtual therapist in providing emotional support

CVE Chatbot | LangChain, Elasticsearch

Built a LangChain-based chatbot that answers vulnerability-related questions from the National Vulnerability Database (NVD) using LLM tool calling for dynamic retrieval and contextual reasoning, improving answer accuracy
Integrated Elasticsearch with hybrid search (dense + sparse retrieval) to improve response relevance for complex cybersecurity questions

DBonK8s | Go, Kubernetes, Terraform, GCP, Kustomize

A chatbot that gives users and admins an easy way to manage databases for development and testing

Built a chatbot to simplify database provisioning (MySQL, PostgreSQL, Redis, MongoDB) for development and testing environments
Utilized Kubernetes API, namespace and RBAC to create secure, multi-tenant resource management

Yungen's blog | Hugo, GitHub Actions

My personal website that automatically syncs with my Obsidian notes through GitHub Actions

Developed a personal website that automatically syncs with Obsidian notes
Implemented a CI/CD pipeline with GitHub Actions for content sync and deployed using Cloudflare Pages

bookmark | TypeScript, Node.js, Hugo, GitHub Actions

My personal bookmark website that automatically syncs with my Notion notes through GitHub Actions

Utilized Notion SDK to export Markdown files from Notion notes and Hugo to generate a static site
Automated the generation, updating, and deployment processes using GitHub Actions, eliminating the need for manual reconfiguration and deployment, resulting in a 30-minute time savings per edit

“Prompt Optimization with Human Annotations Using Pairwise Comparison,” in Proceedings of the IEEE International Conference on e-Business Engineering (ICEBE), 2024, pp. 15–22, doi: 10.1109/ICEBE62490.2024.00012.
“Fusion model using resting neurophysiological data to help mass screening of methamphetamine use disorder,” IEEE Journal of Translational Engineering in Health and Medicine, doi: 10.1109/JTEHM.2024.3522356.