PrepAI: A Web Platform for Placement Preparation with Aptitude Tests, Online Coding Practice, Mistake Review and AI-Generated Study Plans

Abstract

In today\'s challenging job market, effective placement preparation has taken on increased significance, driven by economic instability and heightened instances of layoffs. This makes securing a suitable placement more critical than ever for students and job seekers. However, existing resources often fall short in providing the necessary support and materials. PrepAI is a web-based placement preparation platform that brings aptitude practice, coding practice and AI-generated career guidance into one place. It offers timed aptitude tests, a coding workspace that runs the student\'s code against test cases, a Mistake Notebook that collects failed attempts, a dashboard with progress figures, and JobFit AI, which turns a job description into a day-by-day study plan. The front end uses React with Vite, the back end uses Node.js and Express with JWT login, and the data sits in a PostgreSQL database on Neon. Code is run through Judge0, and the Groq API produces the AI feedback and study plans. This paper explains the design, describes how each module works, and presents the results of the running prototype with screenshots. No study with other students has been run yet, so no measured improvement in scores is claimed.

Introduction

Engineering students preparing for campus placements often use separate platforms for aptitude practice, coding exercises, and interview preparation. These tools generally operate independently and provide limited personalized guidance, making it difficult for students to identify weaknesses, track progress, and develop preparation plans tailored to specific job roles.

This paper presents PrepAI, an integrated AI-powered placement preparation platform designed to address these challenges by combining aptitude testing, coding practice, mistake tracking, performance analysis, and personalized career planning in a single application. The platform uses Large Language Models (LLMs) to analyze job descriptions, provide feedback on coding solutions, identify skill gaps, and generate customized study plans.

The literature review highlights existing developments in AI-based recruitment, automated code evaluation, adaptive testing, competency extraction, and career recommendation systems. These studies demonstrate the potential of AI and machine learning to improve personalized learning and job preparation.

The proposed PrepAI platform consists of five main modules:

1. Aptitude Test: Provides timed multiple-choice tests with 25 randomly selected questions for subjects such as Quantitative Aptitude and Logical Reasoning. It automatically evaluates answers, displays scores, and generates AI-based feedback on strengths and weaknesses.

2. Coding IDE: Allows students to solve programming problems using Python, Java, C++, and JavaScript. The platform uses Judge0 to execute code against test cases and provides results for individual tests. AI-generated feedback suggests improvements to the submitted solutions.

3. Mistake Notebook: Stores unsuccessful coding attempts and allows students to search and filter mistakes by problem title, difficulty, and programming language. This feature helps students revisit previous errors and improve their problem-solving skills.

4. Dashboard: Displays practice statistics, successful submissions, coding pass rates, subject-wise skill levels, recent activity, and focus areas. It helps students monitor their progress and identify topics requiring further practice.

5. JobFit AI: Analyzes a job description, the student's existing skills, and the available preparation time to generate a personalized study plan. It provides a role overview, skill-gap analysis, a daily schedule, recommended learning resources, and relevant interview topics.

The system follows a three-layer architecture. The frontend is developed using React and Vite, while the backend uses Node.js and Express to manage API requests and application logic. PostgreSQL, hosted on Neon, stores user accounts, questions, and submission records. JSON Web Tokens (JWT) provide authentication, Judge0 handles code execution, and the Groq API supports AI-generated feedback and personalized study plans. A shared database allows the platform's modules to use the same student performance records.

The implementation demonstrates how the platform's modules work together through a functional prototype. Sample dashboard results show five practice submissions, two successful submissions, a 50% coding pass rate across four coding attempts, a Logical Reasoning score of 55%, and a Quantitative Aptitude score of 70%. These results illustrate the platform's ability to record performance and identify areas for improvement. However, the reported screenshots demonstrate system functionality rather than measured improvements in student learning or placement success.

Conclusion

PrepAI is a single web platform for placement practice. It combines timed aptitude tests, a coding workspace backed by Judge0, a Mistake Notebook, a dashboard and the JobFit AI planner. Because every attempt goes into one database, the platform can show a student where they are weak and suggest a plan for a specific job. The screenshots in Section VI show that each module works in the prototype. The next steps are: 1) test the platform with a group of students and measure changes in aptitude scores and coding pass rates; 2) grow the question bank and add adaptive question selection, as in [6] and [7]; and 3) add mock interviews and resume analysis, keeping the fairness concerns in [2]-[4] in mind.

References

[1] Saatç?, M., Kaya, R., & Ünlü, R. (2024). Resume screening with natural language processing (NLP). Alphanumeric Journal, 12(2), 121-140. [2] Lo, F. P. W., Qiu, J., Wang, Z., Yu, H., Chen, Y., Zhang, G., & Lo, B. (2025, June). AI hiring with LLMs: A context-aware and explainable multi-agent framework for resume screening. In 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (pp. 4184-4193). IEEE. [3] Kim, C., Choi, J., Yoon, J., Yoo, D., & Lee, W. (2023). Fairness-aware multimodal learning in automatic video interview assessment. IEEE Access, 11, 122677-122693. [4] Bevara, R. V. K., Mannuru, N. R., Karedla, S. P., Lund, B., Xiao, T., Pasem, H., ... & Rupeshkumar, S. (2025). Resume2Vec: Transforming applicant tracking systems with intelligent resume embeddings for precise candidate matching. Electronics, 14(4), 794. [5] Tseng, E. Q., Huang, P. C., Hsu, C., Wu, P. Y., Ku, C. T., & Kang, Y. (2024, December). Codev: An automated grading framework leveraging large language models for consistent and constructive feedback. In 2024 IEEE International Conference on Big Data (BigData) (pp. 5442-5449). IEEE. [6] Hang, C. N., Tan, C. W., & Yu, P. D. (2024). MCQGen: A large language model-driven MCQ generator for personalized learning. IEEE Access, 12, 102261-102273. [7] Yu, J., Zhuang, Y., Sun, Y., Gao, W., Liu, Q., Cheng, M., ... & Chen, E. (2025, July). TestAgent: An adaptive and intelligent expert for human assessment. In Findings of the Association for Computational Linguistics: ACL 2025 (pp. 724-747). [8] Siswipraptini, P. C., Warnars, H. L. H. S., Ramadhan, A., & Budiharto, W. (2024). Personalized career-path recommendation model for information technology students in Indonesia. IEEE Access, 12, 49092-49105.

Copyright

Copyright © 2026 Sudhanshu Bhaurkar, Aasim Ansari, Shlok Kotpalliwar, Dr. Amita Meshram. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.