CrackInterview

CrackInterview.AI is an AI-powered mock interview platform designed to help users practice and improve their performance in tech job interviews. I led the UX/UI design, focusing on creating a streamlined user flow and an intuitive, engaging experience.

Team size: 2
Duration: 8 weeks

My role

As the lead designer on the project, I was responsible for overseeing the entire design process—from research to final delivery. I led the interaction and UX strategy while collaborating with a design intern, who supported visual design tasks such as UI polish, posters, and marketing assets.

  1. User Research & Competitor Analysis: Collaborated with the founder to conduct user interviews, uncovering key user needs and pain points to inform product direction. Analyzed competing platforms to identify strengths, weaknesses, and design opportunities that guided our positioning and feature prioritization.
  2. UX & UI Design: Mapped out user flows, created low-fidelity wireframes, and developed high-fidelity UI designs with a focus on clarity and engagement.
  3. AI Interview Interaction Optimization: Enhanced the AI interviewer's natural language processing capabilities to create more realistic and engaging interview conversations, making the simulation feel more authentic and closer to real interview scenarios.
  4. Usability & Product Testing: Conducted testing sessions to validate design assumptions and gathered user feedback during showcase events to assess product effectiveness.
  5. Iteration: Refined and improved the product design based on testing results, stakeholder feedback, and changing user needs.
CrackInterview Project

Project Background & Goals

CrackInterview.AI was developed as an extension of liba.space, a career platform known for its mentorship network and job training programs for Chinese-speaking professionals. While liba.space provided valuable guidance, users needed a more interactive and scalable way to practice for tech interviews—especially when targeting roles in North America.

To meet this need, we created CrackInterview.AI

Project Background & Goals

Design Process

Our design process followed an iterative, user-centered approach, combining rapid prototyping, AI prompt design, and continuous user feedback.

Research & Discovery

To better understand the challenges faced by early-career job seekers—especially international students targeting tech roles in North America—we conducted a mixed-method research phase combining surveys and semi-structured interviews.

Research Persona

Key Insights

  • Lack of feedback: Most candidates never hear back after interviews and struggle to understand what went wrong.
  • Limited interview opportunities: International students face higher rejection rates due to visa constraints and limited networks.
  • Overwhelming platforms: Tools like LinkedIn and Indeed feel cluttered, repetitive, and not personalized enough for users' specific needs.
  • Desire for targeted guidance: Many users prefer human mentorship and contextual advice over generic AI suggestions.
  • Referral barriers: Getting a referral remains one of the biggest hurdles without insider access or strong alumni networks.

Core Features

Based on our research findings, we designed CrackInterview.AI to address the most pressing needs of job seekers—particularly international candidates navigating the North American job market. The platform blends AI-driven simulation with human support to create a personalized, effective preparation experience.

Smart Job Matching

Smart Job Matching

A custom-built job crawler continuously monitors hiring platforms and, through AI-based matching algorithms, pushes high-relevance job opportunities to users in real time based on their profiles and preferences.

AI-Powered Mock Interviews

AI-Powered Mock Interviews

The system simulates realistic multi-stage interview flows tailored to different roles and scenarios, and generates structured reports with context-specific, actionable guidance for performance improvement.

Mentor Guidance

Mentor Guidance

The platform leverages liba.space's mentor network to offer personalized career coaching based on users' mock interview performance, with targeted guidance on job search, communication, and skills improvement.

Referral Support

Referral Support

The platform enables internal referrals through a trusted network of mentors and verified industry insiders, offering users access to relevant job opportunities once they are interview-ready.

User Flow

User Flow

Conversation Design

Conversation Design

I crafted voice-interactive conversational flows that mimic natural, human-like dialogue to simulate realistic interview experiences tailored to different tech roles and interview stages

AI Interviewer Persona

AI Interviewer Persona 1 AI Interviewer Persona 2

AI interviewers are matched to each interview stage to simulate realistic tone, style, and expectations across the hiring process.

Chatflow

1. Welcome & Setup
AI: "Hi, welcome to your mock interview. I'll be guiding you through a few questions based on your selected role and interview stage."
Let the user know it's a voice-based interview
Encourage a quiet environment
2. Confirm Start
AI: "Are you ready to begin?"
Wait for voice input: "Yes" → proceed
If silent for 3 seconds, give gentle reminder:
"Take your time. Just say 'Yes' when you're ready."
🔹 a. Ask Question
AI: "Tell me about a time you had to resolve a team conflict."
🔹 b. Wait for User Response (Max: ~1 minutes)
🕒 Silence Handling:
After 4 seconds of silence – 1st reminder:
"You can start whenever you're ready."
Still no response – 2nd prompt:
"Think about a project where you disagreed with a teammate."
Still no response – 3rd fallback:
"Would you like to skip this question and move on?"

If skipped → jump to next question.
If user responds → proceed to follow-up.
🔹 C. Response Summary & Transition
AI: "Got it. Sounds like you handled that situation thoughtfully."[Short pause]"Let's move on to the next question."
4. Repeat for Next Questions
Typically 3–5 questions per round
Interviewer persona changes based on selected round
5. End Interview
AI: "Great job today. You stayed thoughtful and composed throughout—that's already a strong skill in any interview." [Short pause]"I'm putting together your report—you'll be able to check it for suggestions on how to improve."
Chatflow

Interactive Demo

Through iterative tuning and testing with ElevenLabs, we enhanced the naturalness and accuracy of the AI interviewer’s responses. As a UX designer, I focused on refining the dialogue flow to create a more realistic and immersive interview simulation experience, ensuring the AI could replicate the tone and rhythm of real interview scenarios.

Interactive Demo

Experience our AI interviewer in action. Click to try the interactive demo and see how natural conversation flows work in practice.

Wireframe

Wireframe

Main Features

Mock Interview

Our AI-powered mock interview simulates real job interviews based users' resume and the selected job description. The entire session is video recorded, helping users reflect on their performance and track progress over time.

Due to the higher cost of running AI digital interviewers, the platform offers two mock interview formats based on the user's subscription plan. Users can freely choose interview modes based on their preferences and subscription level:

  • AI Digital Interviewer Mode: Includes a lifelike AI avatar conducting the interview for a more immersive experience.
  • Standard Mode: A more lightweight version without the digital avatar—ideal for users who prefer self-guided practice while observing their own performance.
AI Digital Interviewer

AI Digital Interviewer

An AI-generated interviewer simulates a lifelike virtual recruiter, enhancing immersion and realism. Questions are asked via voice and displayed with subtitles at the bottom, allowing users to practice in a human-like scenario.

Standard Mode

A simpler text-based interview interface where users can enable their webcam to observe their body language in real-time. Conversation history is shown on the right panel, helping users track the full flow of the mock interview and self-reflect more easily.

Standard Mode
Code Mode

Code Mode

Designed for technical roles, this mode presents real-time coding questions and provides a built-in code editor. Users can type, run, and explain their logic while answering—mimicking real coding interviews.

Mock Interview result

The Result Page provides users with a comprehensive, structured summary of their AI mock interview performance. Designed to be insightful yet easy to navigate, this page helps users understand their strengths, identify weaknesses, and take actionable next steps.

Mock Interview Result

Job Pages

The Job List page intelligently connects users with curated, high-match job opportunities based on their resume and job preferences. Each listing includes a clear match percentage, allowing users to focus on roles that best fit their background.

Job Pages

Future Improvements

To further enhance the platform's value for both job seekers and employers, the following improvements are planned:

Partner Job Integration + Report Sharing

Partner Job Integration + Report Sharing

Collaborate with companies to access up-to-date job openings. Qualified candidates can choose to share their interview reports directly with employers—boosting trust and improving screening success rates.

Talent Pool for Employers

Talent Pool for Employers

Build a centralized talent pool where employers can browse candidates who have completed mock interviews, with performance data and resume highlights readily available.

AI Resume Matching + Auto-Apply

AI Resume Matching + Auto-Apply

Develop a system that automatically matches resumes to verified job openings and submits applications on behalf of the user—streamlining the process and reducing missed opportunities.