I’m putting together a web application that turns the entire interview-practice flow into one seamless experience. First, a visitor picks a technology or domain and chooses the difficulty level. They then authenticate through social media—Facebook, Google or LinkedIn—so a robust, secure social-login flow is essential. Once signed in, the user can upload either a résumé or a job description in PDF or Word format. Your backend should parse the document, identify key skills and context, and instantly generate a tailored set of interview questions. The next step is an AI-powered mock interview, ideally with real-time voice (and, if practical, video) so the system can follow up naturally. After the session finishes, I want a multimodal analysis engine—text, audio and video—to rate performance, uncover sentiment cues, and surface constructive feedback on a dashboard that’s clear and actionable. Deliverables • Fully tested social-login module for Facebook, Google and LinkedIn • Upload component that accepts PDF and Word files and feeds the question generator • LLM-driven question-generation microservice with difficulty control • Interactive mock-interview interface with voice (and optional video) capture • Analytics layer that produces a post-interview feedback dashboard in under five seconds • Clean, well-documented codebase plus deployment scripts and a brief hand-over walkthrough Acceptance criteria: social logins succeed for all three providers, résumé parsing hits at least 90 % accuracy on a standard test set, and the feedback dashboard refreshes promptly after each interview. Feel free to suggest the stack you believe fits best—React, Next.js, Node, Python, LangChain, OpenAI, Whisper, MediaPipe, or anything comparable—as long as it meets the goals above. Looking forward to your approach, timeline, and examples of related AI or interview work.