The Internship Recommendation System helps students and first-time applicants find internship opportunities that best match their skills, education, interests, and location preferences.
Developed for the PM Internship Scheme, the system focuses on accessibility, simplicity, and inclusivity — ensuring even candidates from rural and remote regions can easily identify suitable internships.
Instead of showing a long list, the system suggests 3–5 most relevant internships presented in a clean, card-based interface.
Clean and engaging UI designed for easy navigation and gamified input collection.
- AI-based personalized recommendations using TF-IDF and cosine similarity
- Gamified user input form for collecting:
- Skills
- Education level
- Sector/domain interests
- Location preferences
- Top internship recommendations displayed as cards (title, company, location, details)
- Option to browse all available internships
- Intuitive UI designed for users with low digital literacy
This project follows a content-based filtering approach:
-
Data Preprocessing
Internship descriptions and skills are vectorized using TF-IDF (Term Frequency–Inverse Document Frequency). -
User Profile Encoding
User inputs (skills, interests, education) are combined into a text profile. -
Similarity Calculation
The system uses cosine similarity to measure how closely each internship matches the user’s profile. -
Recommendation Output
The top 3–5 internships with the highest similarity scores are recommended and displayed on the frontend.
| Layer | Technologies |
|---|---|
| Frontend | React.js, Node.js, JavaScript |
| Backend | Python (Flask) |
| Database | MongoDB |
| Libraries Used | scikit-learn, pandas, numpy, Flask |
| Model Techniques | TF-IDF Vectorization, Cosine Similarity |