IntelliStream incubates early-stage systems research across intelligent applications, evolving-data infrastructure, and model execution.

Projects remain in IntelliStream while their abstraction, maintainers, documentation, tests, reproducibility, licensing, and release process are still taking shape. Mature projects graduate into the organization that owns their long-term technical boundary.

IntelliStream research incubator
        | mature projects graduate
        +----------------+-------------------+
        v                v                   v
    RIDE Lab          DataSys            vLLM-HUST
 agent-native         data systems       inference runtime and
 systems research                         hardware execution
        \                |                 /
         +---------------+----------------+
                         v
                       SAGE
              shared flagship product

• RIDE Lab conducts agent-native systems research and stewards the core repositories of SAGE.
• DataSys owns framework-neutral stream, graph, vector/index, online-update, query, lifecycle, and benchmark systems.
• vLLM-HUST is the independent inference substrate and owns model runtime, KV/cache scheduling, compilation, kernels, and hardware execution.
• SAGE — Streaming-Augmented Generative Execution — is the ecosystem's shared flagship product. It applies streaming-computing principles to LLM inference and agent execution; Sage Mate is an application built with SAGE.

These organizations collaborate across distinct technical boundaries. RIDE Lab research systems and products call vLLM-HUST; RIDE is not a runtime layer above it. IntelliStream spans the ecosystem as an incubator rather than acting as another runtime layer.

• DataSys project directory - 15 graduated public data-system repositories.
• Flagships: MorphStream, CANDOR-Bench, GRACE, and Sesame.

• RIDE Lab project directory - agent-native systems research plus the operational home of the shared SAGE core series.
• Product entry points: SAGE, SAGE Docs, SAGE Agentic, SAGE RAG, SAGE Eval, SAGE Studio, and SAGE Tutorials.
• Application: Sage Mate.

Model-runtime, cache-scheduling, compiler, kernel, and hardware-execution projects graduate to vLLM-HUST after maintainer review.

A project is ready to graduate when it has:

• a clear public abstraction and ecosystem boundary;
• named maintainers and durable organization ownership;
• documented setup, supported environments, and contribution path;
• tests and a reproducible evaluation path;
• explicit licensing and citation information where applicable;
• a sustainable release or artifact-maintenance process.

Private, unpublished, or collaborator-sensitive repositories require confidentiality, access, secrets, package, and submission-policy review before transfer.

• sage.org.ai - SAGE product and public documentation.
• ride-lab.github.io - RIDE Lab research organization and portfolio.
• lab.sage.org.ai - IntelliStream team and research portal.
• github.com/intellistream - active incubator projects and migration redirects.

Use the issue tracker and contribution guidance in the relevant project repository. For ownership or graduation questions, contact the project maintainers before moving or renaming code.

• Email: shuhao_zhang at hust.edu.cn
• Website: lab.sage.org.ai