- Deploying and managing Kubernetes-based application- Implementing CI on Jenkins and CD on ArgoCD to apply features on Kubernetes cluster- Developing Helm chart to install declaratively Kubernetes-based application- Using Istio Service Mesh to communicate among services in Kubernetes cluster- Designing the metrics to monitor GPU usage on Prometheus

• Developed 3D Object Detection model to evaluate Object Detection and Tracking models which were integrated into Advanced Driver Assistance System (ADAS)• Implemented semi-automatic 2D object detection and segmentation for Computer Vision Annotation Tool, which saves more than 60% of annotation time when compared to fully manual annotation

• Joined Vingroup AI Engineer Training program to learn core AI subjects: Computer Vision, Natural Language Processing, Basic and Advanced Machine Learning • Implemented Voice-based Gender and Region Classification: classify the audio files spoken by the Vietnamese into 3 region classes (North/ Central/ South) and 2 gender classes (male/ female) with highest accuracy of 78% on the dataset of Zalo AI challenge Voice Gender classification• Implemented Pneumothorax Segmentation: segment X-rays medical image to detect pneumothorax (lungs-related illness) on the dataset provided by Kaggle challenge - SIIM-ACR Pneumothorax Segmentation with accuracy of 89% on top-1 ranking

• Built 2D Object Detection model and integrated the model into Intel embedded devices which were used as Robot eyes

• Built a sentiment analysis in Tensorflow to classify customer feedback