AI-Driven Career Development Platform

Abstract

The process of finding a job is difficult enough, but when searching for professional assistance in the form of interview preparation or career counseling, the cost is usually prohibitive. This is why many individuals find themselves using multiple websites, applications and inputting the same information repeatedly. Therefore, we developed ARIA. ARIA is a completely free online platform that provides both interview practice and career guidance in a single location. ARIA utilizes artificial intelligence (AI) to simulate realistic mock interviews that include audio interaction with the user as well as provide immediate feedback based upon how the user presents him/herself. Additionally, ARIA will provide users with customized career guidance, resume reviews, and a learning plan tailored to each individual user. Users do not need to have any special technology to access ARIA; anyone who has internet can utilize the service regardless of their geographical location or income level. In this paper, we outline the process by which ARIA was designed, explain the technical components of ARIA, and discuss the major ethical issues surrounding AI in career development. Background: The majority of career development products operate independently of each other and therefore do not communicate effectively, limiting a person\'s ability to navigate the career development process. Furthermore, the cost of these products as well as where a person resides limits the availability of effective career development resources. Objectives: We sought to create a product utilizing artificial intelligence that would integrate the interview preparation process and the career development process into one product. Our objectives were to make our product accessible, to ensure the ethics of our product, and to remove traditional barriers to accessing career development resources. Methods: ARIA was created using Next.js 15, React 19, and PostgreSQL on top of a microservice architecture. As for the brain of ARIA, we utilized Google Gemini 2.5 Flash for text-based AI [1] and Hume AI for the voice interview experience as well as the emotional feedback. To maintain accountability and to provide transparency to the users of ARIA, we ensured that humans remained involved in the process. Additionally, we maintained solid security measures to protect the integrity of the data provided to us by users of ARIA. Finally, we established communication between all of the various components of ARIA. Results: ARIA currently offers two primary services—voice-based interview preparation with feedback from multiple angles and chat-based career development resources such as resume review, learning plan, and content creation. The users had a smooth and easy to use experience due to the intelligent feedback they received as a direct result of how everything was integrated. The fact that it is free and online makes it available to anyone who wishes to utilize this resource. Conclusion: ARIA demonstrates the potential of good design to combine several AI technologies and create an environment that supports people rather than hinders them. By utilizing user centered design along with commercial AI technology, we were able to improve access to career resources without sacrificing key issues such as fairness, clarity and user control.

Introduction

Finding a suitable career path has become challenging because individuals often rely on multiple disconnected platforms for job search, interview preparation, and career guidance. This lack of integration forces users to repeatedly input the same information and prevents them from using insights gained in one service to improve another. Additionally, professional career counseling is expensive, often costing $100–$300 per hour, making expert guidance inaccessible to many graduates, career changers, and individuals with limited budgets. Advances in Artificial Intelligence (AI), particularly Large Language Models (LLMs) and conversational AI, offer an opportunity to provide scalable and affordable career assistance.

The research introduces ARIA (AI-Driven Recruitment and Interview Assistant), a career development platform designed to address the shortcomings of existing tools. Current career platforms face three main issues: they operate as isolated systems without data integration, they have limited AI capabilities, and they are often inaccessible due to cost, location, or technical barriers. ARIA aims to solve these problems by connecting multiple services through a unified architecture, using multimodal AI (text, voice, and emotional analysis), and ensuring AI acts as a support tool rather than replacing human decision-making.

Previous research in AI-based career guidance mainly focused on rule-based or machine learning systems that generate recommendations using data such as grades, interests, and activities. While these systems can provide personalized recommendations, they often lack empathy and raise privacy concerns. Similarly, AI interview preparation tools can evaluate verbal and non-verbal responses, but many provide feedback only after the session rather than in real time. Personalized learning systems also exist, using adaptive learning and generative AI to tailor educational content to users’ needs. However, most systems address only one aspect of career development rather than integrating career guidance, interview preparation, and learning support.

ARIA’s system architecture follows four principles: integration of all services, AI assistance with user control, accessibility for anyone with internet access, and delivering immediate value to users. The platform includes two main components: ARIA Prep and ARIA Career Agent. ARIA Prep helps users prepare for interviews through structured practice sessions, AI-generated questions, voice-based responses, emotional feedback, and performance tracking. ARIA Career Agent functions as a virtual assistant that provides career guidance, resume analysis, personalized learning plans, job search strategies, and automated cover letter generation.

Technically, ARIA uses a microservices-based architecture where components operate independently but share core infrastructure. The system uses tools such as Clerk for authentication, PostgreSQL for data storage, and Drizzle ORM for database queries. AI functionality is powered primarily by Google Gemini 2.5 Flash for text analysis and Hume AI for voice interaction and emotion detection. LangChain manages AI conversation flows, while other services handle tasks such as background processing, secure file uploads, and email notifications.

The platform integrates multiple AI tools to perform tasks such as generating interview questions, analyzing resumes, providing career advice, creating learning paths, and simulating interviews with voice-based interaction. Voice simulations analyze speech characteristics such as tone, pitch, speed, and confidence to provide realistic interview feedback. The system stores user interactions in a structured database to allow continuous learning and progress tracking.

Security is a major component of the platform, with OAuth 2.0 authentication, JWT session management, HTTPS encryption, data hashing, and strict input validation. Additional protections such as parameterized queries, content security policies, and rate limiting help prevent common cyber threats.

Key implemented tools include a Resume Analyzer, which evaluates resumes and provides detailed improvement suggestions, and a Roadmap Generator, which creates a personalized learning path toward a user’s career goals. The platform also generates tailored cover letters and stores all interactions in a searchable history.

Conclusion

ARIA provides both Interview Prep and Career Guidance within a single interface, utilizing smart AI and a focus on Accessibility. The manner in which we have designed the architecture of ARIA (utilizing multiple components and providing a single, unified architecture) has resulted in a platform that is stronger than the sum of its individual components. We have utilized a Microservice Architecture along with Shared Infrastructure to allow for Deep Integration of the services, while also allowing for each Service to function Independently as needed. We selected the correct Commercial AI Models to utilize the benefits of Commercial Capabilities, while avoiding becoming tied to One Way of Doing Things.

References

[1] Vaswani, A., et al. \"Attention is All You Need.\" Advances in Neural Information Processing Systems, 2017. [2] Raghavan, M., Barocas, S., Kleinberg, J., & Levy, K. \"Mitigating Bias in Algorithmic Hiring: Evaluating Claims and Practices.\" Proceedings of the Conference on Fairness, Accountability, and Transparency, ACM, 2020. [3] Fernández-Martínez, C., & Fernández, A. \"AI and Recruiting: The Use of Artificial Intelligence in Resume Screening and Candidate Assessment.\" Journal of Business Research, Elsevier, 2020. [4] Cappelli, P., Tambe, P., & Yakubovich, V. \"Artificial Intelligence in Human Resources Management: Challenges and a Path Forward.\" California Management Review, 2019. [5] Naim, I., Tanveer, M. I., Gildea, D., & Hoque, M. E. \"Automated Analysis and Prediction of Job Interview Performance.\" IEEE Transactions on Affective Computing, IEEE, 2018. [6] Campion, M. A., Campion, J. E., & Hudson, J. P. \"Structured Interviewing: A Note on Incremental Validity and Alternative Question Types.\" Journal of Applied Psychology, 1994. [7] Lee, M. K. \"Understanding Perception of Algorithmic Decisions: Fairness, Trust, and Emotion in Response to Algorithmic Management.\" Big Data & Society, 2018. [8] Chien, S. Y., & Chen, S. Y. \"Semantic-Based Personalized Recommendation System for Job Seekers.\" Expert Systems with Applications, Elsevier, 2020. [9] Black, J. S., & van Esch, P. \"AI-Enabled Recruiting: What is it, What is it Becoming, and What Can It Be?\" Journal of Business Research, Elsevier, 2020. [10] Ahmed, S., Hossain, M. S., & Muhammad, G. \"Emotion Recognition from Body Movement.\" IEEE Access, 2021.

Copyright

Copyright © 2026 Dafnia M, Jashera S, Rahul V, Vishnu Priya L, Dr. J. C. Miraclin Joyce Pamila . 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.