Johnson Health Tech is a renowned global provider of innovative fitness and wellness areas. In addition to solid hardware R&D and manufacturing capabilities, software technology is also a key development part, covering the IoT, mobile, web, media streaming, cloud, big data and AI. Recently, we develop and maintain multiple software platforms including @SOUL, @Zone, @ZoneCN, @Mirror, FitDisplay, DataLake, etc. As the leader of the CTT backend team, my role is pivotal in managing a highly skilled team and ensuring the delivery of scalable cloud-based services for our valued customers. Our team is proficient in utilizing Agile tools, Scrum methodologies and embracing DevOps culture to enhance productivity and promote seamless cross-team communication. It is worth noting that I advocate to use AWS services as building blocks, just like Lego pieces, to build our product. Leveraging the cutting-edge of AWS services, including CDK, AppSync, EventBridge, Personalize, Opensearch, Media*, to achieve smooth integration of backend and frontend components and deliver a unified and engaging user experience. Furthermore, I am also interested in exploring the new technologies in our company. In the past year, we spent 20% working time and cooked several POCs of Shopify EC, recommender system, Web3, data lake and GenAI in fitness products . The POC process is interesting and the results are remarkable for us.For the team management, to be a patient and humble mentor, my job is to put everyone in the right place. I conduct monthly 1-on-1 meetings to review work results and set goals with individual team members. Moreover, I am committed to promote a culture of sharing. My team members and I frequently share new technologies within the company. We give talks in the official AWS activities and user group monthly meetings. As advocates of open source principles, we contribute valuable tools and resources back to the community, fostering collaboration and innovation.

I am the lead server engineer in the Whoscall Core team. My job is to develop and maintain phone number related works including the phone number API, AWS cost optimization, data ETL and data research with several machine learning techniques like GNN and NLP. The goal is to provide our users with faster, more accurate caller-id experience. For instance, I am organizing a project which based on Event-Driven Architecture (EDA) and leverages the AWS Lambda@Edge, DynamoDB and EventBridge to speed-up phone number API for our global users, especially in Brazil, recently.Furthermore, we also designed several caller-id prototypes for landline phones based on the Microchip IoT platform. When the landline phone rings, the POC can show the Whoscall caller-id information on the LCD panel and speaks out the caller name by mic. The final product is the Whoscall Wire®. * Below are the tools we used: - Server: Python/uWSGI/Flask, Lua/NGINX/OpenResty, NodeJS/AWS lambda, K8S/EKS - Database: Redis, MySQL, MongoDB, AWS DynamoDB - CI/CD: AWS CDK (Typescript), Jenkins, AWS Code*, Docker - Data: Hadoop, Spark, SparkSQL, Presto, AWS EMR, AWS Glue, AWS Athena, Parquet - Machine Learning: Scikit-learn, Pandas, Tensorflow/Keras, PyTorch, AWS Sagemaker

Pentagon's smart price tracking technology works with customer's products to identify the best pricing opportunities on the Australia E-Commerce markets – giving our customer the advantage and keeping business ahead of their competitors. They will benefited from reduced costs, greater market share, boosted turnover, increased profits and lower supply prices.Our product is a powerful BI dashboard which is a single page application (SPA) based on Backbone. We use Django to construct the RESTful API. For the product data, we use Scrapy framework to crawl competitor's product information and save into Percona MySQL Database. Then, our chief data scientist develops a matching framework based on neighborhood analysis and distance learning algorithm to match the products in a highly efficient way.

We used django to implement a Supply-Chain system for gas station. This includes basic ERP function, caller ID Notification based on Comet, PDF document generation, REST-API for mobile app, etc.

We developed a TWSE stock aggregator by Golang and a distributed stock back-test system based on Gearman to accelerate the stock analysis. BTW, we saved the stock information in MongoDB. MongoDB is so amazing at development stage.

We developed ITRI-ROHC library, PDCP protocol layer, Linux kernel porting, linux network driver and applications on Versatile ARM-926EJS and Versatile ARM11 development board.

We designed and maintained services include mail(sendmail, postfix, spamassassin, virus detection), LDAP, Wireless AAA.