Note for Kiro Judges:

The submission version of this project is preserved on the hackathon-submission branch. The main branch contains active post-hackathon development.

Shadows

AI-powered technical phone screens for software engineers. We handle first-round interviews so your engineering team can focus on building.

Currently offering mock interviews (D2C) to gather feedback on developer experience.

Overview

Shadows conducts live technical interviews with candidates using a conversational AI agent. Candidates solve coding problems while the AI asks clarifying questions, provides hints when appropriate, and evaluates problem-solving approach in real-time.

Tech Stack

• Frontend: Next.js 15, TypeScript, Tailwind CSS
• State Management: Zustand
• Data Fetching: TanStack React Query
• Code Editor: Monaco Editor
• Realtime Voice: WebRTC + OpenAI Realtime API
• Code Execution: Judge0 API
• Database: Supabase (Postgres + pgvector)
• Auth: Supabase Auth with OAuth providers

Getting Started

Prerequisites

• Node.js 18+
• npm or yarn
• Supabase project
• OpenAI API key
• Judge0 API key (RapidAPI)

Installation

npm install

Environment Variables

Copy .env.example to .env and fill in your credentials:

cp .env.example .env

See docs/ENVIRONMENT_VARIABLES.md for detailed configuration.

Development

npm run dev

Open http://localhost:3000 in your browser.

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                         Frontend (Next.js)                       │
│  ┌──────────────┐  ┌──────────────┐  ┌──────────────────────┐   │
│  │ Monaco Editor │  │  Voice I/O   │  │   Session Manager    │   │
│  └──────────────┘  └──────────────┘  └──────────────────────┘   │
└─────────────────────────────────────────────────────────────────┘
                              │
              ┌───────────────┼───────────────┐
              ▼               ▼               ▼
      ┌──────────────┐ ┌──────────────┐ ┌──────────────┐
      │   Realtime   │ │   Judge0     │ │   Supabase   │
      │   Agent      │ │   (Code Exec)│ │   (Storage)  │
      └──────────────┘ └──────────────┘ └──────────────┘

Core Flow

1. Start Session → User selects a problem and begins interview
2. Live Conversation → Realtime agent introduces the problem and engages with candidate
3. Code & Test → Candidate writes code in Monaco, runs tests via Judge0
4. End Session → Session data bundled and sent to evaluator model
5. Scorecard → Detailed feedback with per-dimension scores and evidence

Database Schema

Key tables in Supabase:

• questions - Interview problems with difficulty levels
• starter_codes - Language-specific boilerplate per question
• test_cases - Input/output pairs for validation
• sessions - Interview session state and metadata
• submissions - Code submissions with results

Project Structure

src/
├── app/           # Next.js app router pages
├── auth/          # Authentication utilities
├── components/    # React components
├── hooks/         # Custom React hooks
├── lib/           # Utilities and API clients
├── middleware.ts  # Auth middleware
└── store.ts       # Zustand store

Documentation

• Environment Variables
• OAuth Setup

License

See LICENSE for details.