Mar 2021 - Present
MLops:
• Cooperate with Google Inc to Build MLops on GCP for Customers.
• Build ETL pipeline by using airflow and dataflow on GCP.
• Use Kubeflow Python SDK to build training pipeline with AutoML and Custom
model.
• CI/CD with Cloud Repository and Cloud Build on GCP.
• Continuous train pipeline build on GCP.
ETL:
• Batch Export Cloud SQL table to Cloud Storage by using Cloud Kubernetes Engine
• Build ETL job by using Cloud Workflow on GCP
NLP:
• Build sentence classification model by using BERT, tf-idf, word2vec, Sentence-BERT
in the government agency project.
• Build Docker Image for model serving.
• Use Django framework for RESTful API.
Graph Data Science:
• Use Graph Algorithms in Neo4j to detect smuggling crime in RFP for the
government agency.
• Use Neo4j to build knowledge graphs.
• As an employee training speaker about teaching Neo4j Graph Database for Cathay.
• Applying Data Science in Finance: Neo4j with GDS and ML for Financial Fraud
Detection