EUNJIN KWON

Professional Experience

• Collaborated with the Autonomous Driving and Computer Vision research teams to elevate ML and ETL systems to a service level.
• Created REST APIs to build and maintain Kubeflow pipelines with standardized and reusable components.
• Implemented automation algorithms to remove idle Kubernetes instances occupying GPU nodes in pipeline runtime or post-completion, reducing operational costs.
• Built Apache Airflow to run on NAVER cloud infrastructures and a CI/CD pipeline using Helm Chart and Github Action to deploy on an independent environment per research team.
• Contributed to data pipeline creation and monitoring of multi-cluster operation in Airflow.
• Improved data extraction time by 57% by developing a data serialization application with Apache Spark and documented the implementation and use.

• Focused on the back-end development of Robot Delivery service with Go and MySQL, refactoring the logging module to keep track of robots inside a building through Webhook.
• Lessened the daily parcel non-delivery rate by 13% by developing the notification for parcel managers.

• Created a web application in Python to examine the dialogue level of chatbots through AWS MTurk.
• Developed REST APIs that perform queries on examination results in MySQL to simplify data analysis.

• Increased model accuracy by 0.03% by developing a data cleansing tool and refactoring a data loader.
• Streamlined a model evaluation process by writing a multiprocessing program in Python.

Technical Skills

Education