AI Internship Recommendation Engine

An AI-powered recommendation system that suggests the most suitable internships to students based on their skills, interests, and academic background — built using Collaborative Filtering, Content-Based Filtering, and a Hybrid approach.

Built as part of the Free Online AI & Data Science Internship (Task ID: AI-SS-002) at Data Alcott Systems.

📌 Overview

This project recommends internships to students in three ways:

1. Collaborative Filtering — recommends internships liked by similar students (based on rating patterns).
2. Content-Based Filtering — recommends internships whose required skills closely match a student's own skills, using TF-IDF and cosine similarity.
3. Hybrid Recommendation — combines both approaches (60% collaborative, 40% content-based) for more balanced suggestions.

The project also includes an evaluation module that measures recommendation quality using:

• RMSE (Root Mean Squared Error) — how close predicted ratings are to actual ratings
• Precision@K — of the internships recommended, how many were actually relevant
• Recall@K — of the relevant internships, how many were successfully recommended

🛠️ Technologies Used

• Python
• Pandas
• NumPy
• Scikit-learn (TF-IDF, cosine similarity, train/test split)

📂 Files

File	Description
collaborative_filtering.py	Main script — contains all recommendation logic and evaluation metrics
README.md	This file
PROJECT_REPORT.md	Detailed write-up of approach, results, and limitations

▶️ How to Run

pip install pandas numpy scikit-learn
python collaborative_filtering.py

📊 Sample Output

🎯 Recommendations for Student S001:

📊 Collaborative Filtering:
   - Java Developer (Web) | predicted score: 4.0
   - Data Scientist (AI) | predicted score: 3.5

📊 Content-Based Filtering:
   - AI Intern (AI) | similarity: 0.678
   - Data Scientist (AI) | similarity: 0.556
   - Full Stack Developer (Web) | similarity: 0.098

📊 Hybrid Recommendation:
   - Java Developer (Web) | combined score: 2.4
   - Data Scientist (AI) | combined score: 2.322
   - AI Intern (AI) | combined score: 0.271

📈 Evaluation Report (RMSE / Precision / Recall):
   - RMSE: 1.0
   - Precision@3: 0.167
   - Recall@3: 0.333

⚠️ Limitations

The dataset used is a small, sample dataset (5 students, 5 internships, 9 ratings) built for demonstration purposes. As a result, evaluation metrics can appear noisy — this reflects a real-world challenge in recommendation systems known as the cold-start problem, which is discussed further in the project report.

👤 Author

Shreyashi Khansali — Data Alcott Systems AI & Data Science Internship