Mar 2022 - Present
- Initiate Kubernetes cluster for SaaS in AWS (EKS) for better availability and horizontal scalability by Terraform and Helm.
- Initiate Prometheus, Grafana and Loki for monitoring, alerting and log management system.
- Initiate e2e testing with Cypress.js and integrating new CI/CD pipeline with ArgoCD and github action.
- Improving SaaS availability and horizontal scalability by refactoring legacy system.
- Refactored SaaS RBAC functionality by OPA (open policy agent).
- Refactored the LoRaWAN network architecture and enhanced throughput with AWS IoT Core and new data piepline architecture with Apache Kafka.
- Developed an instance segmentation model for chicken identification using Mask R-CNN and vision transformer model.
- Implemented image processing and feature engineering using binocular stereo vision and chicken recognition outcomes, extracting features of chickens in both 2D and 3D.
- Estimated the weight of each chicken using extracted features with an XGBoost model; the trained model achieved an accuracy rate of 99%.
- Established a data pipeline for a high-throughput real-time chicken image recognition system to predict the weight of chickens on the farm.