VoiceCoach

Job Options Page
Home Page / Interview Prep Page

Inspiration We wanted to transform the stressful, "black box" interview process into a data-driven personal growth journey. Our goal was to build an autonomous coach that senses user confidence and adapts instantly, moving beyond static lists of questions.

What it does VoiceCoach is a multi-modal AI that listens to your confidence (Modulate), watches your presentation (Reka), and researches your target company (Yutori). It maps your skills into a context graph (Neo4j/Pioneer) to provide real-time, personalized interview simulation.

How we built it We integrated five specialized AI APIs into a FastAPI backend and React frontend. By leveraging Neo4j for memory and custom orchestration for tone management, we created a system that evolves its personality based on user performance.

Challenges we ran into Synchronizing five concurrent AI services while maintaining low latency was our biggest hurdle. We also focused heavily on fine-tuning the orchestrator to ensure tone shifts felt natural and helpful rather than robotic.

Accomplishments that we're proud of We successfully built a truly autonomous system where the "difficulty level" is sensed, not selected. Our linked "Decision Trace" in Neo4j allows users to see exactly why the AI provided specific feedback.

What we learned We discovered that sub-lexical cues—like stress and hesitation—provide deeper insight into candidate readiness than text alone. We also mastered multi-agent orchestration across disparate AI platforms.

What's next for VoiceCoach We plan to implement real-time video micro-coaching for body language and expand our knowledge graph to include industry-specific expertise. Global multi-lingual support for international job seekers is also on the roadmap.