TalentCore AI — Autonomous Multi-Agent Corporate Recruitment Pipeline

An enterprise-grade, full-stack recruitment ecosystem designed to eliminate hiring bias and automate deep technical candidate screening. This platform deploys a decoupled, multi-stage asynchronous multi-agent framework built with CrewAI, FastAPI (Python 3.12), and React 18.

---

🏗️ System Architecture & Data Flow

TalentCore AI isolates its heavy multi-agent execution context onto background threads using native Python event pools, preventing user interface freezes or connection thread blockages during token processing loops.

+-----------------------------------------------------------------------+

|                    React 18 / Tailwind CSS Client                     |
+-----------------------------------+-----------------------------------+
                                    |
                                    v  (REST API / OAuth2 Bearer Tokens)
+-----------------------------------+-----------------------------------+

|               FastAPI Async Backend Routing Gateway                   |
+-----------------------------------+-----------------------------------+
                                    |
                                    v  (asyncio.to_thread Event Pools)
+-----------------------------------+-----------------------------------+

|       AGENTIC COMPLIANCE ORCHESTRATION PIPELINE LOGIC ENGINES         |
|                                                                       |
| [ Crew 1: Scanners ] -> [ Crew 2: Examiners ] -> [ Crew 3: Graders ]  |
+-----------------------------------+-----------------------------------+
                                    |
                                    v  (LiteLLM Automated Fallback)
+-----------------------------------+-----------------------------------+

|      Google Gemini API Quota Multi-Key Rotation Management Hub        |
+-----------------------------------------------------------------------+

🧠 Multi-Agent Compliance Pipeline

The backend background worker orchestrates three independent CrewAI agent frameworks to cross-examine technical candidate competencies:

1. Crew 1 (Profile Scanner): Triggered instantly upon PDF resume upload. It parses structural text strings, cross-references them against job requirements, identifies background gaps, and drafts 3 highly specific technical scenario queries.

2. Crew 2 (Cross-Examiner): Engages when the candidate submits their primary answers. It dynamically scans responses for superficial engineering buzzwords, cross-checks them against the initial queries, and generates 2 detailed follow-up probing questions.

3. Crew 3 (HR Grading Compliance Panel): Compiles the final evaluation metrics. It grades the candidate across a strict 4-dimension matrix, generates a structured markdown assessment report, and appends deterministic hiring recommendations based on verbatim textual evidence.

---

🎨 System Workspaces & UI Innovation

• The Candidate Pod: An immersive workspace detailed with a responsive, GPU-Accelerated CSS Vector Processing Orb that dynamically transforms its expressions (Thinking, Talking, Celebrating) in real-time based on background agent calculation milestones.

• Corporate Command Center: An Indigo Deep-Tech administrative workspace featuring a Dual-Pane Operations Hub. It includes independent component scroll lanes to keep tracking matrices securely anchored and features React-Markdown rendering engines that parse raw text tokens into data grids.

• Decision Engine Controls: Interactive header utilities that read the true evaluation scorecard content, automatically displaying an ambient green "Advance to Next Round" click action or a red "Send Rejection Notice" button to log real-time email dispatch traces in the terminal.

📦 Local Workspace Installation & Quickstart

Ensure you have Python >=3.12 and Node.js >=18 installed on your MacBook. This project uses UV for high-speed package management and environment tracking.

1. Configure the Shared Environment (.env)

Create a .env file at your backend root directory and add your unnumbered multi-key Gemini API tokens to support automated rate-limit rotation:

GEMINI_API_KEY_1=AIzaSy...
GEMINI_API_KEY_2=AIzaSy...
GEMINI_API_KEY_3=AIzaSy...

2. Initialize the Backend Server

# Navigate to the backend directory context
cd backend

# Install dependencies and spin up the FastAPI service via UV
uv run uvicorn backend.app:app --reload --host 0.0.0.0 --port 8000

Your backend will initialize a persistent SQLite pool and watch for network handshakes on port 8000.

3. Initialize the Frontend Interface

Open a second terminal window pane and boot your Vite dev server:

# Navigate to the frontend workspace
cd frontend

# Install UI modules and launch the client portal
npm install
npm run dev

Open your web browser and navigate to http://localhost:5173/ to log into your client accounts.

🔐 Administrative Compliance Verification Accounts:

• HR Administrator Desk Login:
  • Email: admin@company.com
  • Password: SecurePassword123
• Candidate Interface Suite: Click register to establish a localized testing profile, upload a text-based PDF resume, and run the complete multi-stage automated consultation pipeline.