Geetanjali Institute of Technical Studies, Dabok , Udaipur (Raj.) Department of Computer Science and Engineering November,2025 A SYNOPSIS of MAJOR PROJECT on AI-Based Internship Recommendation Engine For PM Internship Scheme Submitted by Anmol Parmar(22EGICS010) Alfaiz Sheikh(22EGICS006) Ankit Mod(22EGICS009) Project Guide Head of Department Ms. Harish lakshakar Dr. Mayank Patel Geetanjali Institute of Technical Studies, Dabok , Udaipur (Raj.) Department of Computer Science and Engineering November,2025 Problem Statement: The PM Internship Scheme receives a massive, continuous volume of applications from youth across India, including diverse candidates from rural areas, tribal districts, urban slums, and remote colleges. Many of these applicants are first-generation learners with limited digital exposure and resources. This diversity creates a significant challenge: candidates struggle to navigate the portal and identify which internship opportunities best match their specific skills, interests, or aspirations. This leads to a high volume of misaligned applications and, consequently, missed opportunities for deserving students. Crucially, there is a lack of an intelligent, personalized system that can bridge the gap between the applicant's profile and the available opportunities. The core problem this project addresses is the need for an intuitive recommendation engine that efficiently maps a student's skills and interests to the most relevant internships, ensuring a more equitable and effective selection process. Brief Description: This project proposes the development of an AI-Based Internship Recommendation Engine designed to support the PM Internship Scheme. The system will utilize Natural Language Processing (NLP) and Machine Learning (ML) to analyze candidate profiles (e.g., "skills," "interests," "aspirations") and match them against the requirements of available internship postings from the Ministry of Corporate Affairs (MoCA). The AI engine will provide personalized, relevant recommendations to each applicant. This will simplify the discovery process, particularly for users with limited digital literacy, making the scheme more accessible and increasing the quality and alignment of applications Geetanjali Institute of Technical Studies, Dabok , Udaipur (Raj.) Department of Computer Science and Engineering November,2025 Objective and Scope: Objectives: To develop an NLP module capable of parsing and understanding student profiles, extracting key entities like skills, academic background, and stated interests. To implement a machine learning-based recommendation engine (e.g., content-based or hybrid filtering) to accurately match student profiles with internship descriptions. To create a simple, highly accessible, and intuitive user interface that guides applicants through the profile creation and internship discovery process. Scope: The project will focus on matching student profiles to internship opportunities provided under the PM Internship Scheme by the Ministry of Corporate Affairs. The recommendation model will primarily use applicant-provided data (skills, interests) and the text-based internship descriptions. The user interface will be a web-based portal designed for high accessibility, simplicity, and low-bandwidth compatibility. Methodology: The project will follow an iterative Agile methodology, structured in the following phases: Phase 1: Data Collection and Preprocessing: Gather and anonymize sample applicant profiles and collect a dataset of internship descriptions from the MoCA portal. Define and standardize the key features for matching (e.g., skill taxonomies, interest keywords). Phase 2: NLP & Recommendation Model Development: • Develop an NLP component (using SpaCy or NLTK) to parse resumes and profile text to extract skills, education, and experience. • Train a recommendation model (e.g., using Scikit-learn with TF-IDF and Cosine Similarity for content-based matching) to score the relevance of internships for a given profile. Geetanjali Institute of Technical Studies, Dabok , Udaipur (Raj.) Department of Computer Science and Engineering November,2025 Phase 3: Backend and API Implementation: Develop the core backend logic (e.g., using Flask/Django) to manage user profiles, store internship data (in PostgreSQL/MongoDB), and serve recommendations via a REST API. Phase 4: Interface Development and Integration: Build the front-end user interface (e.g., using React) focusing on simplicity and accessibility. Integrate the front-end with the backend API to display personalized recommendations. Phase 5: Testing and Evaluation: Perform comprehensive testing (see Section 6) and evaluate the system's recommendation relevance. Hardware and Software Requirements: Hardware: Processor: Intel Core i5 (8th Gen or higher) or equivalent. RAM: Minimum 8 GB (16 GB recommended for model training). Storage: 500 GB HDD/SSD. Software: Operating System: Windows 10/11 or Linux (Ubuntu). Programming Language: Python 3.8+. Database: PostgreSQL or MongoDB (for user profiles and internship data). Web Framework: Flask/Django (Backend) and React/Vue.js (Frontend). Technologies: Core AI/NLP: Scikit-learn (for ML models), Pandas, NumPy, NLTK, SpaCy (for text processing). ML Framework: TensorFlow/Keras or PyTorch (for potential deep learning extensions). Backend: Python (Flask/Django). Frontend/Interface: HTML5, CSS3, JavaScript, React/Vue.js. Geetanjali Institute of Technical Studies, Dabok , Udaipur (Raj.) Department of Computer Science and Engineering November,2025 Testing Techniques: Unit Testing: Testing individual components like the NLP profile parser, the matching algorithm, and API endpoints using Python's unittest or pytest Integration Testing: Verifying the seamless flow of data from the user profile creation (frontend) through the backend API, into the recommendation engine, and back to the user's dashboard. User Acceptance Testing (UAT): Testing the system with a target group of students, especially those representing the "first-generation learner" demographic, to assess usability, clarity, and the perceived quality of the recommendations. Accuracy/Performance Metrics: Evaluating the recommendation model using offline metrics like Precision@k , Recall@k , and NDCG (Normalized Discounted Cumulative Gain) to measure the relevance of the top-k recommendations. Project Contribution: Democratizes Opportunity: This project directly addresses the "digital divide" identified in the problem statement by providing an intelligent, easy-to-use tool for first-generation learners and students from rural/tribal backgrounds, making the PM Internship Scheme more equitable. Improves Application Quality: By guiding students to roles they are genuinely qualified for and interested in, the system will significantly reduce the number of misaligned applications. This saves time and administrative overhead for both the students and the MoCA. Provides Personalized Guidance: The engine acts as a "virtual career assistant," helping students discover opportunities they might have otherwise overlooked, thereby broadening their horizons and improving their chances of success. Efficient Talent-Opportunity Matching: For the Ministry, this tool ensures that the right talent is channeled to the right opportunities, maximizing the effectiveness and impact of the entire internship program.