DEPLOYMENT.md

Steps AI Mock Interview Platform - Deployment & Verification Guide

This guide details the step-by-step instructions to deploy, configure, and verify the Steps AI Mock Interview Platform on production environments (Render for the FastAPI backend, and Vercel for the Next.js frontend).

---

🧠 Backend Deployment (Render)

We leverage Render's native Blueprints infrastructure which reads render.yaml to spin up both a high-performance PostgreSQL database and a FastAPI Python web service in unison.

Step-by-Step Instructions:

1. Prepare and Push Code: Ensure all changes are committed and pushed to your GitHub repository.
2. Go to Render Console: Navigate to render.com and sign in.
3. Launch Blueprint:
   • Click New (top-right) and select Blueprint.
   • Connect your GitHub repository containing the project.
4. Deploy Service:
   • Render automatically parses render.yaml and displays the configured components (stepsai-backend web service and stepsai-db PostgreSQL instance).
   • In the Specified configurations fields, input the following environment variables:
     • GROQ_API_KEY: Paste your API credential key obtained from console.groq.com.
     • API_KEY: Input stepsai_hackathon_2026_key (used to secure endpoints via headers).
     • CORS_ORIGINS: Set as http://localhost:3000,https://stepsai-smoky.vercel.app (binds local dev and your Vercel frontend URL).
   • Click Apply / Deploy.
5. Get Your Backend live URL:
   • Once compilation and database initialization finishes, copy your generated web service URL from the dashboard.
   • Target Live URL Format: https://stepsai-backend.onrender.com

---

🌐 Frontend Deployment (Vercel)

The Next.js premium космический dashboard is optimized for Vercel's edge network, building in seconds using our pre-packaged vercel.json instructions.

Step-by-Step Instructions:

1. Go to Vercel Console: Sign in at vercel.com.
2. Import Repo: Click Add New -> Project and import your GitHub repository.
3. Configure Project Settings:
   • Root Directory: MUST set this to frontend/ (since our frontend code lives in a monorepo subfolder).
   • Framework Preset: Auto-detects as Next.js.
   • Build Command: npm run build (falls back to vercel.json).
4. Define Environment Variable:
   • Under Environment Variables, add the following key-value pair:
     • Key: NEXT_PUBLIC_API_URL
     • Value: https://stepsai-backend.onrender.com (paste your copied Render live URL here).
5. Deploy: Click Deploy. Vercel will build, optimize static chunks, and route it to edge servers in seconds.
   • Target Live URL Format: https://stepsai-smoky.vercel.app

---

🚦 Post-Deployment Alignments (CORS & Security Keys)

1. CORS Update on Render:
   • Now that your Vercel frontend URL is live, go to your Render Web Service dashboard -> Environment page.
   • Ensure CORS_ORIGINS includes your official Vercel domain: CORS_ORIGINS=http://localhost:3000,https://stepsai-smoky.vercel.app
   • Save changes. Render applies the CORS rules instantly.
2. Security Gates Validation:
   • REQUIRE_AUTH is pre-configured to true on the Render web service environment to lock down endpoints.
   • All REST requests must pass the X-API-Key: stepsai_hackathon_2026_key header credential.

---

🧪 Post-Deployment Endpoint Verification (curl command examples)

You can verify that your production services are active, connected, and fully secure by executing these standard curl commands directly from your local terminal.

1. Complete Database Health Connection Probe

Assert that the backend connects successfully to the persistent PostgreSQL instance:

curl -X GET https://stepsai-backend.onrender.com/health

Expected Response (HTTP 200 OK):

{
  "status": "healthy",
  "app_name": "Steps AI Mock Interview Platform",
  "database": {
    "status": "healthy",
    "error": null
  }
}

---

2. Lightweight Uptime check

Assert that the FastAPI event loops are active and track system uptime:

curl -X GET https://stepsai-backend.onrender.com/api/v1/health

Expected Response (HTTP 200 OK):

{
  "status": "healthy",
  "app_name": "Steps AI Mock Interview Platform",
  "environment": "production",
  "uptime_seconds": 1824.5
}

---

3. Ingestion & Resume Intelligence Upload (Locking Auth)

Simulate a PDF resume upload. This command attaches the required X-API-Key auth headers and uploads a file chunk:

# Note: Replace "/path/to/resume.pdf" with any valid PDF or DOCX file path on your local computer
curl -X POST https://stepsai-backend.onrender.com/api/v1/resume/upload \
  -H "X-API-Key: stepsai_hackathon_2026_key" \
  -F "file=@/path/to/resume.pdf"

Expected Response (HTTP 200 OK):

{
  "session_id": "9b1deb4d-3b7d-4bad-9bdd-2b0d7b3d4bad",
  "profile": {
    "name": "Jane Doe",
    "job_role": "Backend Engineer",
    "experience_years": 4,
    "seniority": "Mid",
    "skills": ["Python", "FastAPI", "SQL", "Docker"],
    "summary": "Experienced engineer specializing in distributed REST interfaces and database migrations.",
    "skill_gap_analysis": ["Git", "Algorithms", "Testing"]
  }
}

---

4. Mock Interview Session Launch

Launch a state-aware mock session using the parsed resume token:

# Note: Replace "9b1deb4d-3b7d-4bad-9bdd-2b0d7b3d4bad" with the session_id obtained from the previous upload response
curl -X POST https://stepsai-backend.onrender.com/api/v1/interview/start \
  -H "X-API-Key: stepsai_hackathon_2026_key" \
  -H "Content-Type: application/json" \
  -d '{
    "resume_session_id": "9b1deb4d-3b7d-4bad-9bdd-2b0d7b3d4bad",
    "interview_mode": "Technical",
    "total_questions": 5
  }'

Expected Response (HTTP 200 OK):

{
  "interview_session_id": "a1b2c3d4-e5f6-7a8b-9c0d-1e2f3a4b5c6d",
  "resume_session_id": "9b1deb4d-3b7d-4bad-9bdd-2b0d7b3d4bad",
  "first_question": "Hello Jane! Welcome to our technical interview. I see you have 4 years of experience as a Backend Engineer with expertise in Python and FastAPI. To kick things off, could you explain the differences between relational database indexing and simple document lookup in your past database migrations?",
  "interview_mode": "Technical",
  "difficulty_level": "Mid",
  "total_questions": 5
}