Dana — Multi-Agent Conversational AI Platform
Writeup in progress. Full case study covering DAG orchestration design, operator taxonomy, SSE streaming implementation, human-in-the-loop gates, and production operations coming soon.
All projectsProject
Dana — Multi-Agent Conversational AI Platform
Production DAG-based agent orchestration runtime with 41 domain operators, 20+ tools, SSE streaming, multi-provider LLM execution, JSON Schema validation, and human-in-the-loop gates — serving CanApply's educational guidance workflows.
agentsconversational-aillmproduction
Problem
Educational guidance at scale requires structured conversations spanning admissions, programs, funding, and immigration — with consistent grounding, tool use, and auditability across thousands of concurrent sessions.
Solution
A DAG-based agent orchestration runtime with 41 domain operators, 20+ integrated tools, SSE streaming, multi-provider LLM execution via llm-client, JSON Schema output validation, and human-in-the-loop gates for high-stakes decisions.
Impact
Production conversational AI system serving CanApply users across the full guidance lifecycle: program discovery, admissions prep, funding, and immigration.
Stack
FastAPIllm-clientKafkaPostgreSQLRedisNext.jsPython