Stackular ATS

An internal Applicant Tracking System built for Stackular. Two portals — Recruitment and Interviewer — covering the full hiring pipeline from job description creation to candidate selection.

---

Overview

Stackular ATS streamlines the recruiting workflow: recruiters create AI-generated job descriptions, upload resume folders, and manage candidates through a Kanban pipeline with automated ATS match scoring. Interviewers get a separate portal to view assigned JDs, schedule interviews, and submit feedback.

---

Tech Stack

Layer	Technology
Frontend	React 19, TypeScript (strict), Vite, Tailwind CSS 4
Routing	TanStack Router
Data Fetching	TanStack Query v5
UI	Lucide icons, Sora font
Backend	.NET 10, SQL Server (Azure)
AI / Scoring	Python 3, FastAPI (localhost:8000)
Storage	Azure Blob Storage

---

Features

Recruitment Portal

• JD Creation — AI-generated job descriptions with live field-summary panel (Live Summary), edit, assign, and finalize flow
• Resume Upload — Bulk upload with folder drag-and-drop, duplicate detection via file hashing
• ATS Scoring — Queue-based background scoring pipeline; real-time score updates via SSE stream
• Candidate Pipeline — 10-stage Kanban board (under_review → selected / rejected) with virtualized list view
• Job Folder → Candidate navigation — "View full profile" preserves stage context; back button restores the exact kanban column; prev/next arrows iterate candidates within the same stage
• Dashboard — Pipeline funnel, attention items, application volume chart, recent job folders
• Interviewer Assignment — Assign interviewers to candidates directly from the candidate detail page

Interviewer Portal

• Dashboard — Upcoming interviews, personal stats, recent activity, prep tips
• JD View — Read assigned job descriptions and fill out role descriptions
• Interview Feedback — Submit ratings, recommendations, and notes per interview round

---

Project Structure

ATS-Tool-Stackular/
├── Front-End/          # React + Vite app
│   └── src/
│       ├── features/
│       │   ├── recruiter/      # candidates, dashboard, jd-creation, job-folders
│       │   └── interviewer/    # dashboard, jd-creation
│       └── shared/             # components, hooks, context, types, api client
├── Back-End/
│   └── ATStool/        # .NET 10 Web API
│       ├── Controllers/
│       ├── Services/
│       ├── Models/
│       ├── DTOs/
│       └── Migrations/
└── AI_Models/          # Python FastAPI service
    ├── ats_scoring/
    ├── jd_creation/
    └── resume_extraction/

---

Getting Started

Prerequisites

• Node.js 20+
• .NET 10 SDK
• Python 3.11+
• SQL Server (local or Azure)
• Azure Storage account (for resume blobs)

Frontend

cd Front-End
npm install
npm run dev

Runs at http://localhost:5173.

Backend

cd Back-End/ATStool
dotnet restore
dotnet run

Runs at http://localhost:5122.

AI Service

cd AI_Models
pip install -r requirements.txt
python api_server.py

Runs at http://localhost:8000.

---

ATS Scoring Architecture

Scoring runs asynchronously via a Channel<AtsScoreJob> queue processed by AtsScoreWorker (a .NET BackgroundService). Results are pushed to the frontend in real time over a persistent SSE connection at GET /api/candidates/stream?jobId=.

Score sentinels:

• null — unscored
• -1 — pending (shown as ...)
• negative (other) — failed
• 0–100 — match percentage

---

Portals & Accounts

Role	Route
Recruiter	/recruitment/...
Interviewer	/interviewer/...

Authentication is JWT-based. Roles: Recruiter, Interviewer.

---

Design System

• Theme: Dark (#0c0c0c background), Sora font throughout
• Card pattern: Double-bezel — outer frosted shell + inner #161719 panel
• Color language: Green = active/success · Amber = pending · Blue = selected/AI · Purple = filtered · Red = error/rejected
• Sidebar: 240px desktop · 64px icon-only tablet · hidden mobile