AIInterviewCoach
Link to open source: https://github.com/basti456/AI-Interview-Coach
đź’ˇ Inspiration & The Problem
Preparing for job interviews is high-stakes, stressful, and often lacks realistic, low-latency practice. Traditional mock interview platforms either rely on clunky text-based bots or suffer from sluggish audio lag, destroying the natural flow of conversation. Candidates struggle to find personalized practice tailored to their exact resume and target job description—and rarely get actionable, structured feedback afterward.
🚀 The Solution: What It Does
Interview Coach AI is a native Android application that delivers a hyper-realistic, real-time voice interview experience. It bridges the gap between candidates and recruiters by pairing a modern Kotlin & Jetpack Compose frontend with a robust Python FastAPI backend, powered by Agora's cutting-edge Conversational AI Engine.
Candidates upload their resume and job description, step into a live voice session where an AI interviewer conducts a dynamic interview, and receive a comprehensive post-interview evaluation (scores, strengths, and areas for improvement).
⚡ How Agora Powers This Project
Agora is at the heart of our real-time interactive experience:
•Agora Conversational AI Engine: Powers managed voice inference and natural turn-taking without requiring complex manual LLM streaming infrastructure.
•Agora RTC & RTM SDKs: Deliver crystal-clear, ultra-low-latency voice streaming and signaling between the mobile client and the AI agent backend.
•Seamless Session Management: The Python FastAPI backend orchestrates Agora channels securely, handling token generation and agent connection lifecycles instantly.
✨ Key Features
1.🗣️ Real-Time Voice Conversation: Fluid, natural voice interaction with sub-second latency for a true conversational flow.
2.đź“„ Resume & Job Description Grounding: Tailors every question specifically to the candidate's background and target role.
3.đź’¬ Live Transcript & State Management: Real-time speaker-labeled transcription bubbles so users can visually follow the conversation.
4.đź“Š Post-Interview Evaluation & Scoring: Leverages Groq Cloud (llama-3.3-70b-versatile) to generate a detailed score breakdown, strengths, and actionable feedback immediately upon session completion.
🛠️ Tech Stack
•Android Frontend: Kotlin, Jetpack Compose (Material 3), Coroutines, Retrofit/OkHttp, Agora RTC & RTM SDKs.
•Backend & AI Orchestration: Python, FastAPI, Agora Conversational AI Engine.
•Evaluation Engine: Groq Cloud LLM API.
This build was uploaded as a hackathon project
