Career Profile
7+ years Backend engineering experience
5.5+ years DevOps CI/CD experience
5+ years Data collection, Data cleaning, and Data analysis experience
4+ years Web applications (Rails) experience
3+ years Cloud services experience
1+ years Android native applications experience
6+ months Hybrid applications (React Native) experience
3+ months iOS native applications experience
Experiences
TeamT5 Inc.
Unlike a Team Lead, this role involves not only assisting in talent recruitment for the company but also determining the technical direction for the product, focusing on solving large-scale technical issues, expanding the scope of impact, and conducting code reviews.
- Full-stack development & Maintenance, 2023 - Now
Responsible for the entire backend architecture, including the post-processing, agent engine, product platform, and license system.
ThreatSonar: The platform is an advanced threat hunting and MDR (Managed Detection and Response) solution with a wide range of capabilities. These include compromise assessment, auto-investigation, and orchestration. With over 100 built-in IoCs (Indicators of Compromise) and thousands of backdoor indicators, it enables the discovery of threats and remote execution of incident responses in a post-compromised environment.
- Full-stack development & Maintenance, 2019 - 2023
Features development : Managed the license system, ensuring product functionality through encryption and decryption. Led a small team in developing user interfaces and backend services using Python and Rails.
Backend development : Designed RESTful APIs, WebSockets, and other services (Linux kernel) as standalone systems (on-premises) to build an open virtualization format. Refined existing APIs and imported Swagger to conserve development resources.
Automation, monitoring & troubleshooting : Created comprehensive GitLab CI/CD pipelines for fully automated deployment, improving PR stability. Developed an alarm system with Chatbot and SMTP integration to monitor license and system status for maintenance purposes.
- Architecture planning, 2021 - 2023
Cluster system : Pioneered the implementation of a MySQL, PostgreSQL, and Elasticsearch cluster, leading to significant enhancements in data traffic, replication, and system scalability.
High availability : Achieved a high availability architecture (Keepalived, Active/Standby, GlusterFS, and cluster) to ensure stability and enable failover/failback capabilities.
- System performance, 2020 - 2023
Data migration : Migrated data from MySQL to PostgreSQL within a month, resulting in a 30-50% improvement in complex query performance.
Performance optimization : Refactored the legacy worker dispatcher and implemented profiling with stress testing to enhance system efficiency by 10x.
Cache system : Devised a real-time cache system using Redis with TTL (Time To Live) to reduce response time based on data usage.
Table partition : Designed monthly rotated table partitions to alleviate the database's load, particularly when handling extensive data volumes.
- Malware detection, 2020
Data pipeline & analytics : Built data pipelines using Docker containers with Kibana and Nginx for analyzing malicious behaviors.
Machine learning : Conducted research on anomaly detection and NLP (Natural Language Processing) to intelligently detect threat actors using obfuscated command lines.
D8AI Holdings Corp.
D8-a11y(TSL): It’s a hybrid application (React Native) designed to assist hearing-impaired individuals in communicating via text while at brick-and-mortar stores using a telephone.
- Backend development, 2018 – 2019
Videotelephony & speech transcription : Adopted the Janus framework (WebRTC and SIP) in conjunction with GCP (Google Cloud Platform) Speech to ensure stable streaming media capabilities.
Registration & login system : Built a user registration/login system using MongoDB and JWT (JSON Web Tokens) for efficient management of user sessions and profiles.
Message broker system : Established a comprehensive message broker system (ELK stack, Kafka MQ, and AWS) for centralized log management and data visualization.
Shanghai Droi Technology Co., Ltd
InfoHub: It’s a native mobile application for iOS and Android that focuses on real-time data within the data industry, providing information such as news, videos, funny images, and media retrieval.
AdZodiac: It’s a fully integrated advertising platform based on web services and MoPub mobile SDK, incorporating various advertising APIs such as Facebook, AdMob, Flurry, and more.
- Mobile full-stack development, Launcher, and Advertising SDK, 2016 - 2018
Android deployment & launcher : Developed Android native prototypes and implemented media retrieval from the internal BaaS (Backend as a Service) platform. Integrated Ads into the Android launcher, positioning it on the left swipe home screen.
Data collection, cleaning, analysis : Expanded the data crawler's capabilities by implementing various third-party APIs (such as YouTube, Vimeo, Tenor, Giphy, Makeagif, Gfycat, etc.) as scrapers to obtain external information. Implemented social meta tag parsing and utilized NLP/Boilerpipe information extraction algorithms to extract useful information, including data images and contents, from over 8K data sources, processing over 100K data daily.
Cloud messaging : Devised a notification system using Firebase and BaaS to efficiently deliver 2.5M messages per month for tracking MAU (Monthly Active Users).
Advertising SDK : Assisted an advertising team in developing native applications for Android/iOS platforms and built an Ads SDK for clients, resulting in increased revenues.
- System performance, Infrastructure, and Recommendation, 2017 - 2018
Performance tuning : Designed a parallel processing data crawler with MongoDB cache, resulting in a 24x performance improvement. Compressed over 100K data from more than 8K data sources using FFmpeg, and uploaded them to CDN (Azure CDN with Azure Blob Storage, BelugaCDN with Google Cloud Storage, and Cloudflare) to significantly reduce overall bandwidth costs by over 90%.
Monitoring & troubleshooting : Implemented a comprehensive range of system troubleshooting and monitoring services, including Chatbots on WeChat and Telegram, Webhook integration, SonarQube for code analysis, Systemd and Upstart for service management, Crontab for scheduling tasks, Netdata for real-time monitoring, JenkinsCI for continuous integration, and Docker for containerization. These services proactively prevent system crashes and enable automatic broadcasting.
Recommendation system : Built a personalized news recommendation system utilizing collaborative filtering with news categories and user classes to enhance exposure. Implemented a hybrid approach that combines Word2Vec/Doc2Vec feature-based document representation with SVD/MF model-based methods to improve efficiency.
Education
National Taiwan University
Master's Thesis: Predicting Web Services with Service Ontology
Lab & Advisor: Software Engineering Laboratory, Jonathan Lee, Ph.D
Relevant Research Domains: Prediction and Recommender System, Ontology, Natural Language Processing, Information Retrieval and Extraction, Machine Learning
Relevant Courseworks: Software Engineering, Software Engineering Design (Design Pattern), Computer Security, Interactive Computer Graphics, Computer Vision, Natural Language Processing, Information Retrieval and Extraction, Personal Communication Service
National Central University
Senior Projects: System Detection by Face Orientation, English Learning Companion Robot
Lab & Advisor: Computational Intelligence and Human Computer Interaction Laboratory, Mu-Chun Su, Ph.D
Relevant Courseworks: The Attack and Defense of Computers, Object Oriented Analysis and Design, Computer Animation and 3D Games, Virtual Reality, Multimedia Signal Processing, Software Hardware Codesign, Computer Networking
Societies: Service Club of NCU (Member, Sep 2009 - Jun 2011), Sign Language Club of NCU (Member, Sep 2009 - Jun 2010), Student Association of NCU CSIE (Member of IT Department, Sep 2009 - Jun 2010)
Awards & Honors
Served as an invited speaker for achieving outstanding performance amongst deaf people.
Ranked second place in 2009 Taiwan College Entrance Examination in sciences major (with biology) for hearing impaired students.
Got admitted to TOP 3 graduate schools(NTU, NTHU, NCTU) in Taiwan by achieving general student entry requirements.
Imagine Cup
Designed a system using C# and Microsoft Kinect, combined with CILAB’s robot, to assist children in learning English. The system incorporates interactive features, such as imitating sounds and actions, to enhance the learning experience.
Advanced to the Final Round
Subject : English Learning Companion Robot
Members : 3
Innovation Competition
Built a system using C# that enables parents to monitor their children’s learning progress. The system retrieves image information through Kinect for analysis. All the collected data can be uploaded to a network platform for parents to review and download.
Advanced to the Final Round
Subject : System Detection by Face Orientation
Members : 3
Third Edition Competitive Programming of Robotic Car
Utilized the CI-STAMP evaluation board (provided by CILab) to develop a self-propelled vehicle program using C. The program was designed to navigate and overcome challenges in a maze.
Won the Championship (More than 20 teams)
Members : 4 (Team Leader)
Achieved over a semester grade point average of A(4/4.3).
