How It Works

MAESTRO operates as a web-based platform readily accessible to public officers via their GSIB machines, ensuring secure data exploitation.

Since August 2023, MAESTRO has integrated end-to-end production capabilities for Generative AI and Machine Learning Operations onto its platform, enabling agencies to efficiently scale the development and deployment of AI/ML models.

The tools/services on MAESTRO include:

• Pre-defined configurations and templates based on MLOps best practices so that pipeline setups can be accelerated.
• Access to expanding selection of generative AI models like Stable Diffusion and LLMs from model repositories such as HuggingFace, as well as those from various Cloud Solution Providers such as AWS, Azure and GCP. These models are readily available for fine tuning and eventually deployment via API Gateway to serve agencies’ production systems.
• Scalable high-compute resources (i.e. GPU-as-a-Service) to support computationally-intensive analytics projects efficiently.
• Up-to-date code libraries/packages for both R & Python.
• Means for code sharing and collaboration amongst users both within and across agencies.
• No-code machine learning tool to generate predictions without needing to write any code.