AI Interview Coach

Interface

🔥 𝗜𝗻𝘀𝗽𝗶𝗿𝗮𝘁𝗶𝗼𝗻 Identified a common “Placement Gap” among students Students have technical knowledge but struggle with: Interview pressure Communication & articulation Idea: Build a “Interview Simulator” Safe environment Practice + fail + improve before real interviews

⚙️ 𝗪𝗵𝗮𝘁 𝗜𝘁 𝗗𝗼𝗲𝘀 (𝗔𝗜 𝗜𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄 𝗖𝗼𝗮𝗰𝗵) Resume Parsing Extracts skills using AI Generates personalized technical questions Simultaneous Monitoring Uses OpenCV for eye tracking & focus detection Uses microphone for voice capture Real-time Grading Evaluates answers instantly using Llama 3.3-70B Provides feedback in < 2 seconds Audit Dashboard Generates Cyber-Dark analytics report Includes: Hireability Index 360° competency mapping

🏗️ 𝗛𝗼𝘄 𝗪𝗲 𝗕𝘂𝗶𝗹𝘁 𝗜𝘁 Architecture Modular State-Machine Design Frontend Built with Streamlit Custom enterprise-grade CSS AI Layer Integrated Groq API for fast inference Uses Llama 3.3 models Parallel Processing Python Multithreading Handles: Camera (OpenCV) Audio recording simultaneously Data Visualization Plotly → Radar charts Pandas → Behavioral trend analysis

⚠️ 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗙𝗮𝗰𝗲𝗱 Hardware Synchronization Issue Webcam + Microphone conflict (“Resource Busy”) Solution Used background threading for audio Main thread handled vision processing Prompt Engineering Ensured AI stays in “Recruiter Mode” Maintained resume context throughout interview

🏆 𝗞𝗲𝘆 𝗔𝗰𝗰𝗼𝗺𝗽𝗹𝗶𝘀𝗵𝗺𝗲𝗻𝘁𝘀 Built Radar Chart for soft skills Confidence Fluency Focus Achieved Near Zero-Latency Feedback Real-time evaluation Developed Integrated Proctoring System “Red-Box” alerts for distraction detection

📚 𝗪𝗵𝗮𝘁 𝗪𝗲 𝗟𝗲𝗮𝗿𝗻𝗲𝗱 Multimodal AI = Future of EdTech Combines Vision + Voice + Text Learned Weighted Scoring Systems Final Grade Formula: 𝐺 = 𝜔1⋅Focus+𝜔2⋅Fluency+𝜔3⋅ATS Match