ZeroChômage

LOGO

🚀 What inspired us The idea behind ZeroChômage was born from the frustration many job seekers face: navigating interviews, language barriers, and repetitive application processes. We imagined a world where AI could assist, represent, and empower candidates by simulating interviews, adapting to different roles, and automating job submissions.

🧠 What we learned We learned how to: Build a conversational AI using LLMs that can simulate human interviews Integrate Speech-to-Text (Whisper) and Text-to-Speech (ElevenLabs) for live voice interaction Animate a 3D avatar using Three.js and Ready Player Me Use LangChain and Pinecone for context memory and conversation history Design a multilingual, inclusive system (English, German, French)

🛠️ How we built it

Frontend:     Next.js, Three.js, WebRTC
Backend:      Python, Node.js, Docker
AI:           GPT-4, Whisper, ElevenLabs
Databases:    Firebase, Pinecone
Tools:        LangChain, Ready Player Me, REST API

We built a modular system that supports: Mock interview mode for preparation Live interview assistant that answers on your behalf Job application automation tailored to your CV

⚠️ Challenges we faced Real-time avatar-lip sync with audio responses Handling multilingual interactions dynamically Maintaining conversational context with long interviews Creating human-like responses while staying concise Ensuring fast performance in low-bandwidth environments