🧠 TalentScout AI

Smart Resume Screening Platform (AI/ML)

---

🚀 Overview

TalentScout AI is an AI-powered resume screening platform that analyzes resumes and maps candidate skills to job requirements using semantic understanding and skill inference, instead of traditional keyword matching.

Conventional ATS systems often reject capable candidates simply because exact keywords are missing. TalentScout AI solves this by understanding meaning, context, and implied skills, making recruitment fairer and more intelligent.

“An ATS that thinks like a human recruiter.”

---

📸 Screenshots

Job Database & Dashboard	Smart Resume Analysis
Manage open roles and track applicants	See semantic scores and missing skills

---

✨ Key Features

• 📄 Universal Input: Resume upload support (PDF, DOCX, TXT)
• 🧠 Semantic Analysis: Contextual understanding using transformer embeddings.
• 🛠️ Smart Skill Inference: Automatically credits implied skills.
  • Example: Resume says "Django" → System infers "Python".
• 📊 Hybrid Match Scoring: 60% Semantic Context + 40% Hard Skill Match.
• 🔍 Gap Analysis: Clearly lists missing skills with improvement suggestions.
• 🏆 Auto-Ranking: Recommends the best job role across multiple openings.
• 🖥️ Professional UI: Clean, intuitive interface built with Streamlit.

---

🧠 How It Works

1️⃣ Input

• HR adds one or more Job Descriptions to the database.
• Candidate uploads a Resume (PDF/DOCX).

2️⃣ Semantic Understanding

• Both texts are converted into vector embeddings using a Sentence Transformer (MiniLM).
• Cosine Similarity measures the contextual alignment between the resume and the job description.

3️⃣ Smart Skill Mapping

• Skills are extracted using:
  • Keyword detection (Regex/NLP)
  • Inference logic (Framework → Core Skill mapping)
• The system identifies matching skills and critical missing skills.

4️⃣ Scoring Logic

The final match percentage is calculated using a weighted formula:

Final Score = (Semantic Similarity × 0.6) + (Skill Match Score × 0.4)

5️⃣ Output

• Match percentage
• Recommended job role
• Missing skills list
• Comparison table across all open roles

---

📊 Why TalentScout AI?

Feature	❌ Traditional ATS	✅ TalentScout AI
Matching Method	Keyword-based	Semantic (Meaning-based)
Skill Inference	❌ No	✅ Yes (Infers Context)
Missing Keywords	Auto-reject	Context-aware analysis
Transparency	Black-box	Clear gap analysis
Role Comparison	Manual	Automatic ranking

---

🛠️ Tech Stack

• Language: Python
• AI/ML: Sentence Transformers (all-MiniLM-L6-v2)
• NLP: Cosine Similarity, Regex
• File Handling: PyPDF2, python-docx
• Frontend/UI: Streamlit
• Data Handling: Pandas

---

▶️ How to Run Locally

1. Clone the Repository

   git clone [https://github.com/Vani691/TalentScout-AI.git](https://github.com/Vani691/TalentScout-AI.git)
   cd TalentScout-AI
   Create Virtual Environment
   

Bash python -m venv venv

Windows:

venv\Scripts\activate

Mac/Linux:

source venv/bin/activate Install Dependencies

Bash pip install streamlit sentence-transformers torch PyPDF2 python-docx pandas Run the App

Bash streamlit run app.py 🎯 Target Users HR Teams: To automate initial screening.

Recruiters: To find "hidden gems" rejected by keyword filters.

Hiring Managers: To quickly rank candidates.

Placement Cells: To help students optimize resumes.

👥 Team – Code Alchemists Project built for Multiverse of Tech Hackathon 2026 at Smt. Indira Gandhi College of Engineering.

Shravani Mane (Team Lead)

Shubham

Sanika

Sunaina

📌 Hackathon Note This project was developed during the official hackathon window as a working MVP. Open-source AI models were used, while all application logic, scoring strategy, UI flow, and skill inference mechanisms were implemented by the team.