PrepGenius AI

Parallel Interview Practice System For Data Careers

One session. Structured AI evaluation. Persistent weak-area intelligence. Ready-to-present report cards.

Data Analyst • Data Scientist • AI/ML Engineer

4 Interview Modes • 8 Questions Per Session • Real-Time Scoring • Readiness Dashboard

Edunet Foundation × IBM SkillsBuild Internship
Employability Skills & AI Track • 2025

---

Quick Access

Overview • Screenshots • Tech Stack • Architecture • Workflow • How It Works • Quick Start • Environment • Troubleshooting • Roadmap

Why This Matters: AI interview orchestration, progressive evaluation, semantic weak-area memory, and product-grade UX in one portfolio project.

---

Overview

PrepGenius AI is an interview preparation platform for data-focused roles. Instead of one-off question generators, it provides a complete closed loop:

1. Generate targeted questions for role, mode, and difficulty.
2. Evaluate each answer with structured, actionable feedback.
3. Persist attempts and weak areas across sessions.
4. Produce a report card with a practical 4-week action plan.

What Makes It Different

Traditional Practice Tools	PrepGenius AI
Generic random questions	Role-specific and mode-specific question generation
Simple pass/fail scoring	Rich scoring with strong points, gaps, and ideal answers
No continuity between sessions	localStorage-backed attempt memory and trend tracking
Text-only outputs	Full report card, readiness dashboard, and visual breakdowns
Weak user feedback loops	Confidence calibration plus per-question guidance

Core Dimensions

Every session combines multiple product layers:

Dimension	Engine	Output
Question Generation	useAiCoach + Groq	8 tailored interview questions
Answer Evaluation	useGroqEvaluator + Groq	Score, strengths, misses, tips, model answer
Progress Memory	useWeakAreaTracker	Topic stats, trends, weak-area prioritization
Readiness Intelligence	Dashboard calculators	Category scores and predicted pass rate

Screenshots

Splash Screen

Landing Experience

Mock Interview Workspace

Workflow Snapshot

Workflow Diagram

flowchart LR
  A[Set Role + Mode + Difficulty] --> B[Generate 8 Questions]
  B --> C[Answer With Confidence Rating]
  C --> D[AI Evaluation With Groq]
  D --> E[Live Feedback + Score]
  E --> F[Store Attempt Metrics]
  F --> G{More Questions?}
  G -- Yes --> C
  G -- No --> H[Generate Final Report]
  H --> I[Weak-Area Dashboard]
  I --> J[Readiness Snapshot + Study Plan]

Loading

System Architecture

flowchart TD
  U[User] --> UI[React UI Layer]
  UI --> APP[App State Machine]

  APP --> COACH[useAiCoach]
  APP --> EVAL[useGroqEvaluator]
  APP --> TRACK[useWeakAreaTracker]

  COACH --> GROQ[(Groq API)]
  EVAL --> GROQ

  TRACK --> LS[(localStorage)]
  LS --> WA[WeakAreaDashboard]
  LS --> RD[ReadinessDashboard]

  APP --> RPT[ReportDisplay]
  RPT --> MD[MarkdownRenderer]

Loading

How It Works

1. User selects role, interview mode, difficulty, and skill focus.
2. useAiCoach calls Groq to generate session-specific questions.
3. User answers each question and optionally sets confidence level.
4. useGroqEvaluator scores responses and returns structured feedback.
5. useWeakAreaTracker persists attempts and updates topic analytics.
6. On final question, a complete report card is generated and rendered.
7. Dashboard views summarize weak areas and readiness snapshot for next practice.

Tech Stack

Application Stack

Layer	Technology	Purpose
Frontend	React 19 + React DOM	Component-based app and UI rendering
Language	TypeScript 5	Type-safe business logic and component contracts
Build Tool	Vite 6	Fast local development and production bundling
Styling	Tailwind utilities + custom CSS	Product UI, motion, glassmorphism, gradients
AI SDK	groq-sdk	Question generation and answer evaluation
Persistence	localStorage	Session continuity and weak-area memory

Runtime Scripts

Script	Command	Description
Dev	npm run dev	Start development server at localhost:3000
Build	npm run build	Produce optimized production assets
Preview	npm run preview	Run built application locally

Key Product Advantages

Capability	Value	Differentiator
Structured AI Evaluation	Better learning per question	Not just score, but guidance and ideal answers
Confidence Calibration	Measures self-awareness	Compare confidence with actual performance
Weak-Area Intelligence	Focused upskilling path	Topic-level tracking across sessions
Visual Readiness View	Fast interview preparedness check	Category heatmap and pass-rate estimate
End-to-End Session Flow	Cohesive user journey	Splash, setup, interview, report, dashboard

Project Structure

PrepGenius-AI/
├── App.tsx
├── index.tsx
├── index.html
├── types.ts
├── data.ts
├── components/
│   ├── SplashScreen.tsx
│   ├── LandingPage.tsx
│   ├── InterviewScreen.tsx
│   ├── PreferencesForm.tsx
│   ├── ReportDisplay.tsx
│   ├── WeakAreaDashboard.tsx
│   ├── ReadinessDashboard.tsx
│   └── ...
├── hooks/
│   ├── useAiCoach.ts
│   ├── useGroqEvaluator.ts
│   └── useWeakAreaTracker.ts
└── Public/

Quick Start

1) Install

git clone https://github.com/Yashaswini-V21/PrepGenius-AI.git
cd PrepGenius-AI
npm install

2) Configure Environment

# Windows
copy .env.example .env

# macOS/Linux
cp .env.example .env

Set your API key:

GROQ_API_KEY=your_groq_api_key_here

3) Run Application

npm run dev

Open: http://localhost:3000

4) Production Check

npm run build
npm run preview

Environment Variables

Variable	Required	Default	Description
GROQ_API_KEY	Yes	None	Groq API key for generation and evaluation

Troubleshooting

Issue	Likely Cause	Fix
Questions not generated	Missing or invalid API key	Re-check .env, restart dev server
Empty/weak response quality	Vague input answer	Use clearer, specific answers with examples
Dashboard seems empty	No attempt data yet	Complete one full session first
Build fails unexpectedly	Dependency mismatch	Run npm install then npm run build

API And Logic Notes

• Question generation and report generation run in useAiCoach.
• Per-answer evaluation runs in useGroqEvaluator.
• Attempt tracking and readiness calculations run in useWeakAreaTracker.
• Markdown report content is sanitized before rendering.

Roadmap

• PDF export from report page
• Voice response mode
• More role templates
• Optional backend relay for protected key usage
• Team and mentor review mode

License

MIT

---

PrepGenius AI
Built for high-clarity mock interview practice and measurable improvement.

Project Links
Repository • Groq • Vite • React

Maintainer
Yashaswini V • GitHub

Internship Collaboration
Edunet Foundation × IBM SkillsBuild Internship · Employability Skills & AI Track · 2025

If this project helps you, please consider starring the repository 🌟 !!