AI Practice Hub
A collection of hands-on AI demos, built to learn by doing.
Hybrid Multi-Agent system with Reactive and Deliberative paths. Queries are classified, complex investigations decomposed by a Planner Agent into parallel sub-tasks, executed with traceable ReAct loops, and reviewed by a Review Agent.
Enterprise KPI dashboard with charted metric trends, telemetry correlation, and AI-generated anomaly diagnosis plus action plans, with demo time anchored to June 8, 2026.
A CloudNova onboarding mentor with self-evolving long-term memory. It remembers the learner, consolidates candidate observations into stable rules, and forgets stale ones — the write → consolidate → promote → recall → forget lifecycle, visualized.
Grounded Q&A over indexed documentation and Git repositories. Compare RAG answers against baseline model responses with source citations, retrieval diagnostics, and multi-provider chat support.
A self-hosted model workspace: inspect available vLLM models, start a runtime on demand, run repeatable evaluations, and monitor live inference performance (TTFT, tokens/sec, GPU cache). Preview is public; live access requires owner authorization or an active invitation.
Preview is available to everyone. Live access requires owner authorization or an active invitation.