Enterprise Architecture · HR Tech
PythonFastAPILangGraphMilvusRedisSSE Stream
Enterprise Interview Agent System: Standardized Talent Evaluation Hub
Designed and built by Behelitai, this talent evaluation hub leverages FastAPI and LangGraph to standardize candidate profiling, structured questioning, and auditable scoring into deterministic, secure workflows.
Interactive Interview Simulator Sandbox
Select a position model and launch the workflow to observe LangGraph's multi-turn reasoning and automated report compilation in real time.
Select Position Competency Model:
LangGraph State Tracker (AssessmentState)
Agent Pipeline Stage:
Ready (Waiting)
Active LangGraph Node:
IDLE
Competency Vector Scores (FastAPI Backend)
AI Orchestration0%
Concurrency & Async0%
Vector Search & RAG0%
Security & Isolation0%
Identified Risk Flags
No risks flagged yet
SSE Streaming Developer Logs Console
DISCONNECTED
AI Interview Window
Interaction input disabled. Click start to begin.
LangGraph Eight-Node Pipeline Routing
Moving away from simple stateless chat, the system utilizes LangGraph to run a tightly-constrained stateful machine.
Node01
1. Load Profile
load_candidate_profile
Node02
2. Load JD Competency
load_position_profile
Node03
3. Milvus RAG Lookup
retrieve_assessment_context
Node04
4. Generate Question
generate_question
Node05
5. Structurally Audit Answer
parse_candidate_answer
Node06
6. Branch Decision Node
decide_next_step (Evaluates loop break)
Node07
7. Trigger Follow-up Qs
generate_follow_up
Node08
8. Compile Report
generate_final_report