Multi-Cloud
Human ICT
AI Vision Management System MSA Development
This project was undertaken to optimize the core AI model performance of Human ICT's Vision-based hazard detection system and to overhaul the infrastructure for maximum scalability and maintenance efficiency.
TOSKY successfully transitioned from the existing Monolithic architecture to MSA (Microservices Architecture), enabling independent operation of each service.
Technically, Kubernetes (k8s) deployment processes were introduced for the Python-based ML components and communication servers, enabling flexible orchestration in cloud environments.
Additionally, a Multi-Tenant architecture was applied to the Java-based management site, allowing a single instance to serve multiple clients, dramatically improving operational efficiency.
Through this project, TOSKY demonstrated advanced AI model optimization capabilities as well as system architecture design expertise leveraging the latest cloud-native technology stack.
Key Features
01
Cloud-Native k8s Deployment Environment
Applied Kubernetes to ML models and communication servers for flexible scaling based on service load and stable deployment processes
02
MSA Architecture Migration & Optimization
Redesigned the existing Monolithic system into a microservices structure to minimize inter-service interference and enhance overall system availability
03
Multi-Tenant Management System Implementation
Applied multi-tenant architecture to the Java-based management site, enabling independent and efficient management of data across multiple organizations