IntelliView Orchestrator: Enterprise Interview & Risk Scoring Platform
Scalable backend microservice engine orchestrating automated interview workflows, configurable risk weight scoring, and asynchronous notifications.
Interactive Client-Side Demonstration • Showcases API schemas, dynamic weight calculation, and evaluation algorithms.
What real-world challenge does this system address?
System Design & Approach
- Dynamic Risk Weight Configurator: Full CRUD API with validation rules ensuring weight matrices sum to mathematical unity.
- Notification Deduplication & Template Engine: Asynchronous event pipeline preventing duplicate alerts across candidate interview lifecycles.
- Asynchronous Worker Architecture: Offloads long-running candidate evaluation tasks to Celery workers backed by Redis.
- Automated Test Validation: Verified through automated Pytest suites and validation schemas.
System Architecture & Dataflow
Client/Recruiter UI -> FastAPI Microservices -> Pydantic Validation -> Celery Background Workers (Redis) -> PostgreSQL (Alembic Migrations)
Exposes endpoints for risk weight management, candidate evaluation, and interview status querying.
Enforces non-negative weight bounds and dynamic fallback weights when position-specific configs are omitted.
Dispatches email digestion, report generation, and interview notification deduplication asynchronously.
Stores position configs and candidate interview logs with versioned database migration scripts.
Stack Specification & Components
Python 3.11
FastAPI & Celery
PostgreSQL (with in-memory store fallbacks for isolated testing)
Docker Compose multi-container stack
Risk Weight Matrix Configurator
FastAPI validation rule simulation: weights must normalize to exactly 100%
{
"position_name": "Senior AI Systems Engineer",
"technical_weight": 0.5,
"domain_weight": 0.3,
"behavioral_weight": 0.2,
"is_normalized": true
}Engineering Decisions
Key design choices, constraints, and solutions formulated during development
Celery Background Workers for Notifications & Scored Reports
Decoupled interview evaluation and notification dispatch into Celery asynchronous tasks.
Kept API response times sub-50ms by never blocking the HTTP request thread with template rendering or external SMTP handshakes.
Automated Fallback Architecture for Position Risk Weights
Implemented hierarchical weight resolution (`custom position -> departmental default -> global system default`).
Prevents scoring pipeline crashes when a newly created job requisition does not yet have custom recruiter risk thresholds assigned.