The problem
Mock interviews are valuable, but they’re hard to get on demand and harder to get honest feedback on. Generic question banks miss the role; well-meaning friends are too kind. The result: most candidates show up to the real thing under-rehearsed and over-stressed.
The product
Interview Ace runs a realistic mock interview over voice. It picks questions that match the role, asks follow-ups, and scores each answer against a per-role rubric. After the session, the recording becomes a study plan — the questions you fumbled, the patterns the AI noticed, and a focused list of what to drill before the real interview.
How it works
- Pick a role. Software, product, consulting, design, finance — each role comes with its own bank of question types and a scoring rubric.
- Live mock. A voice-driven session where the AI plays the interviewer, asks follow-ups, and adapts difficulty as you go.
- Per-answer scoring. Each answer is scored on the rubric dimensions (clarity, structure, evidence, technical accuracy) with concrete examples of where points were lost.
- Study plan. The session output is a focused list of weak spots and the resources to address each one.
Stack
- iOS: SwiftUI with native audio capture.
- Realtime voice: OpenAI Realtime API for the live interviewer; transcription and scoring via the same model post-session.
- Backend: lightweight session/transcript service for storage and replay.
- Distribution: App Store with StoreKit 2 for the paid tier.
Status
Interview Ace is in development.