PrepSphere: A Multi-Agent AI System for Intelligent Placement Preparation and Adaptive Mock Assessment

Preparation for getting a job is usually done using different things like tools to build a resume website to apply for jobs, databases with interview questions videos to watch and tests to practice. Learners might do a lot of things without knowing if they are helpful for a job or if they actually learned something. This paper talks about PrepSphere, a system that brings together checking learning tailored to each person checking skills and simulated tests in one continuous process. The resume analysis gives a score that's good for the systems that companies use to sort resumes. Then the system checks the skills. Compares them with what the job and the company need. These requirements are then turned into plans. Just reading learning materials doesn't mean you learned something: you have to get least 80% on a quiz before your skill record is updated. Then four different agents check your knowledge, your thinking abilities, your answers to HR questions and your communication through typing or speaking. Large language models are used to do things like creating plans suggesting questions and giving feedback. Checking, rules for scoring and how ready you are still done with clear steps. The system uses React and TypeScript, FastAPI, MongoDB, Groq-hosted processing, Piper for speech whisper for transcribing Rhubarb for face movements and a Three.js avatar. The Placement Readiness Score is based on the resume quality the skills you have the interview results and your education. Testing the version of the system shows that it has a good structure but there are still some issues with how it works and how voices are handled that need fixing before it can be used in schools. The method shows how having agents can help prepare for a job without having a hidden system decide if you are ready.