Nguyen Duy Dat

Project Description: Built and deployed a system for automatic Vietnamese poetry generation using apretrained GPT-neo model. ◦ Implementation Steps: ∗ Data Collection: Utilized Selenium to gather data from thivien.net and saved it into a CSV file containing titles, content, and authors. ∗ Data Processing and Model Training: Fine-tuned a pre-trained GPT-neo model using the Hugging Face framework based on the collected data. ∗ API Development: Developed an API using FastAPI, enabling users to send requests and receive generated poetry. ∗ Containerization with Docker: Packaged the application into Docker containers for simplified deployment and management. ∗ Deployment on Kubernetes (K8s): Deployed the application on GKE using Helm templates with 3 replicas, ensuring availability and scalability. ∗ Monitoring and CI/CD: Implemented monitoring systems such as Grafana, ELK, and Prometheus to track system performance. Established CI/CD pipelines on GCE to automate the evelopment and deploy;