Inspiration
What it does
How we built it
Challenges we ran into
Accomplishments that we're proud of
Inspiration
Job seekers practice alone with no real feedback, enter interviews unprepared, and have no way to simulate real interview pressure.
What it does
GemLive Interview is an AI-powered mock interview coach using multi-agent orchestration. 5 specialized AI agents work together: Interviewer conducts realistic questions, Feedback Coach analyzes answers, Resume Analyzer reviews documehttps://gemlive-interview-coach.web.app/nts via Vision API, Performance Coach tracks progress, and Orchestrator coordinates everything.
How we built it
Built with vanilla JavaScript, CSS3, and HTML5 - zero dependencies. Uses two Gemini API patterns: Streaming SSE for real-time conversations and Direct API for analysis. Deployed on Firebase Hosting. The entire app is just 3 files totaling 64KB.
Challenges we ran into
Implementing real-time streaming with Server-Sent Events required careful parsing of chunked responses. Maintaining conversation context across 5 agents while keeping each agent persona consistent was complex. Building a premium UI with glassmorphism and animations without any CSS framework.
Accomplishments that we are proud of
Zero-dependency architecture (3 files, 64KB total), real-time streaming that feels natural, autonomous multi-agent orchestration that preserves context.
What we learned
How to leverage Gemini 2.5 Flash streaming for real-time AI interactions, multi-agent system design patterns, and building premium UIs without frameworks.
What is next
Voice-to-voice interviews with Gemini Live API, industry-specific templates, interview recording and playback, collaborative mock interviews with peers.
Log in or sign up for Devpost to join the conversation.