Hsin-tien, New Taipei City, Taiwan
Oversaw Data Pipeline Infrastructure and Guideline project: :
.Integrated Airflow and BigQuery to provide infrastructure for data pipeline and MLOps
.Built data pipeline to collect business data and statistics from 30k daily records
.Integrated data pipeline into website backend, reduced loading time by 90%
Managed API Serving Infrastructure and Guideline project:
.Built and maintained infrastructure for API serving
.Migrated team’s API servers from separated GCP VMs to GCP Kubernetes Cluster
.Configured DNS and TLS security service for team’s API services
Led Vision-based AI Wound Assessment API:
.Implemented object detection algorithms (Yolo v3) to detect bedsores occurred on elders
.Built CNN model (Inception Resnet v2) to assess bedsores, reached 80% accuracy
.Developed size measurement algorithm to estimate dimensions of bedsores
Created AI products to aid 4,000+ nursing practitioners in over 180 elder-care centers for support decision making
Optimized API by multi-threading computation, saved over 50% of computing resources
Mentored 3 interns in AI and DS related projects