Declarative Data Services: Structured Agentic Discovery for Composing Data Systems
Researchers have introduced Declarative Data Services (DDS), a new architecture designed to improve how AI agents discover and compose data systems. Unlike previous methods that struggled with heterogeneous search spaces and verification, DDS uses a structured approach with typed contracts to break down complex discovery tasks into manageable sub-searches. This framework aims to enable agents to reliably build functional data stacks from declarative user intents, as demonstrated on a trading-backend workload where it achieved convergence where other methods failed. AI
IMPACT This structured approach to data system composition could enable more reliable and efficient AI-driven data infrastructure development.