InitAI — AI Interview Simulator

A full-stack mock interview platform that simulates a real technical interview end-to-end: behavioral screening with a voice AI interviewer, a live coding challenge, and a personalized debrief report.

How it works

1. Intake — describe your target role and upload your resume (PDF or .txt). Claude generates a personalised interview plan calibrated to your level (Intern / Junior / Mid / Senior).
2. Behavioral round — voice-only conversation with "Alex", the AI interviewer. No typing — you speak, Alex listens and responds.
3. Coding round — Monaco editor with a problem matched to your level from the Supabase problem bank. You can ask Alex for hints via voice at any time.
4. Debrief — structured report with category scores, per-category feedback, and a prioritised improvement plan.

Tech stack

Layer	Tech
Frontend	Next.js 16, React 19, TypeScript, Tailwind CSS 4, Monaco Editor
Backend	Spring Boot 3.5, Java 21, Spring Data JPA, Apache PDFBox
AI	Anthropic Claude (claude-sonnet-4-6) via REST
Database	Supabase (PostgreSQL)
Voice	Web Speech API (browser-native, Chrome/Edge)

Project structure

InitAI/
├── backend/          # Spring Boot API
│   └── src/main/java/com/initai/backend/
│       ├── controller/   # REST endpoints
│       ├── model/        # JPA entities + request/response DTOs
│       ├── repository/   # Spring Data repositories
│       └── service/      # Claude integration, session logic, prompts
├── frontend/         # Next.js app
│   └── app/
│       ├── page.tsx                         # Intake form
│       ├── interview/[sessionId]/           # Behavioral chat
│       ├── interview/[sessionId]/code/      # Coding editor
│       └── interview/[sessionId]/debrief/   # Score report
├── dev.py            # One-command dev launcher
└── .env              # Local secrets (not committed)

API endpoints

Method	Path	Description
POST	/api/sessions/plan	Generate interview plan from goal + resume
POST	/api/sessions	Create session from confirmed plan
POST	/api/sessions/{id}/behavioral	Send behavioral message, get Alex's reply
GET	/api/sessions/{id}/problem	Get the assigned coding problem
POST	/api/sessions/{id}/hint	Ask Alex for a hint during coding
POST	/api/sessions/{id}/submit	Submit code, trigger debrief
GET	/api/sessions/{id}/debrief	Retrieve stored debrief report
POST	/api/resume/extract	Extract text from a PDF upload

Local setup

Prerequisites

• Java 21+
• Node.js 18+
• Python 3 (for the dev launcher)
• A Supabase project with the schema below
• An Anthropic API key

Environment variables

Create a .env file in the project root:

SUPABASE_URL=jdbc:postgresql://<host>:5432/postgres
SUPABASE_USER=postgres.<project-ref>
SUPABASE_PASSWORD=your-password
ANTHROPIC_API_KEY=sk-ant-...

Database schema

Run in the Supabase SQL editor:

CREATE TABLE problems (
  id          BIGSERIAL PRIMARY KEY,
  title       TEXT NOT NULL,
  description TEXT NOT NULL,
  constraints TEXT[],
  difficulty  TEXT NOT NULL,
  starter_code TEXT
);

CREATE TABLE sessions (
  id                   TEXT PRIMARY KEY,
  level                TEXT,
  plan_summary         TEXT,
  resume_text          TEXT,
  phase                TEXT,
  problem_id           BIGINT REFERENCES problems(id),
  hint_count           INT DEFAULT 0,
  conversation_history JSONB,
  debrief_report       JSONB
);

Run

python dev.py

This starts both servers and loads .env automatically.

Service	URL
Frontend	http://localhost:3000
Backend	http://localhost:8080

Or start them separately:

# Terminal 1
cd backend && ./mvnw spring-boot:run

# Terminal 2
cd frontend && npm run dev

Voice support requires Chrome or Edge — the Web Speech API is not available in Firefox or Safari.