CareerVivid iOS

Voice-first, personalized mock interview practice that compounds daily.

Comment

CareerVivid iOS — AI interview practice that compounds

CareerVivid turns interview prep into a short, repeatable daily practice loop. Instead of practicing one generic question and losing the feedback, candidates build a visible record of what improved, what still needs work, and which role-specific challenge to tackle next.

The problem

Interview preparation is fragmented. Candidates collect question lists, practice alone, receive vague feedback, then start over the next day. They rarely know whether they are improving or how to connect individual answers to the role they want.

What I built

CareerVivid is a native SwiftUI iOS app for focused mock-interview practice:

  1. Personalized Skill Tree — users choose a target role, experience level, existing skills, and the skills they want to develop. CareerVivid turns that profile into a six-step challenge path.
  2. Company interview quests — real company interview themes and stages from techinterview.org become structured practice loops. Coding and system-design rounds open the specialized web workspaces; recruiter screen, behavioral, values, and final-round practice stay native in iOS.
  3. Voice-first answer loop — tap to record, review an editable transcript, receive concise “strengthen this answer” suggestions, then choose Send for Deep AI Analysis.
  4. Actionable report history — every attempt is saved independently, including multiple attempts at the same question. Reports show communication, confidence, answer relevance, what went well, and the one highest-leverage next action.
  5. Daily momentum — Home visualizes interview activity, active days, streaks, recent scores, and the next practice focus.

How it works

Choose role and skills
  → receive a personalized challenge
  → record or type an answer
  → review the editable transcript
  → get AI analysis and targeted feedback
  → retry or advance with report history preserved

The production voice flow uses Firebase Cloud Functions in us-west1 and Gemini 3.1 Flash-Lite to turn recordings into editable text and to generate focused feedback. The app uses the same canonical interview-question catalog as the CareerVivid web experience, so the company, stage, and question context remain aligned across surfaces.

Built with Codex and GPT-5.6

I had three years of web-development experience but no Swift/iOS background. I used GPT-5.6 and Codex in ChatGPT to translate an existing CareerVivid web workflow into a native iOS product: SwiftUI information architecture, reusable interview/report components, microphone and transcript states, skill-tree personalization, Firebase integration boundaries, QA fixes, and documentation. Codex accelerated implementation while I drove the product requirements, testing, UX iterations, and integration decisions.

What makes it different

CareerVivid is not another isolated mock interview. It connects a candidate’s target role, the skills they want to build, a concrete company-style question, an editable voice response, and durable feedback history into one practice loop that is short enough to repeat daily.

Judge resources

Privacy

Users control the answer sent for analysis: they can edit the transcript before sending it. The app’s privacy center explains audio handling, data retention, export, and account deletion.

Demo video

A public demo video will be added before final submission.

Built With

  • codex
  • firebase
  • gemini-3.1-flash-lite
  • google-cloud-functions
  • gpt-5.6
  • swift
  • swiftui
Share this project:

Updates