DEV Community: Subashbelina

How I Built CVpilotX — An AI Resume Builder with Next.js and OpenAI

Subashbelina — Tue, 16 Jun 2026 11:31:27 +0000

Resumes are generic. Jobs are specific.
Every job needs a tailored CV but nobody
has time to rewrite it every time.

So I built CVpilotX — paste a job description,
AI rewrites your CV to match it, download as PDF instantly.

🔗 Live: https://cvpilotx-frontend.onrender.com/

Stack

Next.js (Frontend)
Node.js + Express (Backend)
OpenAI GPT-4
Docker (Deployment)
Render (Hosting)

Biggest Challenge — PDF Generation

Generating a clean, formatted PDF from
AI-generated content was harder than expected.

The AI returns text but PDF needs proper
formatting — fonts, spacing, margins, sections.

First attempt was ugly. 😅

Fix using puppeteer:

const browser = await puppeteer.launch()
const page = await browser.newPage()
await page.setContent(htmlContent)
const pdf = await page.pdf({
  format: 'A4',
  margin: {
    top: '20px',
    bottom: '20px',
    left: '20px',
    right: '20px'
  }
})
await browser.close()

Convert AI response to HTML first,
then Puppeteer renders it as perfect PDF.
Clean output every time.

What It Does

📋 Paste any job description
🤖 AI rewrites your CV to match it
📄 Download as PDF instantly
⚡ Takes less than 30 seconds

What I Learned

PDF generation needs HTML as middle layer
Puppeteer inside Docker needs extra config
OpenAI prompt engineering matters more than the code itself

Found this helpful? Drop a ❤️

How I Built an AI Writer SaaS App with React and OpenAI

Subashbelina — Mon, 15 Jun 2026 20:50:09 +0000

I wanted to build something useful with AI.
So I built an AI Writer — a SaaS app that generates
blogs, emails, and content using OpenAI GPT.

🔗 Live: https://ai-writer-saas-sepia.vercel.app/
⚙️ Backend: https://ai-writer-server.onrender.com/

Stack

React.js (Frontend)
Node.js + Express (Backend)
OpenAI API
Vercel (Frontend) + Render (Backend)

Biggest Challenge — Render Cold Start

My backend was hosted on Render free tier.
Every time nobody used the app for 15 minutes,
Render would spin down the server.

Next request would take 30-60 seconds to respond.
Users thought the app was broken. 😅

Fix:

setInterval(() => {
  fetch('https://ai-writer-server.onrender.com/ping')
    .then(() => console.log('Server kept alive'))
}, 14 * 60 * 1000)

Added a simple /ping endpoint in Express:

app.get('/ping', (req, res) => {
  res.json({ status: 'alive' })
})

Cold start problem solved. App feels instant now.

Key Features

📝 Blog post generator
📧 Email writer
💡 Content ideas generator
⚡ Instant AI responses

What I Learned

Render free tier spins down after inactivity
Keep-alive ping is the easiest fix
OpenAI streaming makes UX feel much faster

Found this helpful? Drop a ❤️

How I Built CineMood — A Movie Discovery App

Subashbelina — Mon, 15 Jun 2026 20:43:19 +0000

I was tired of scrolling Netflix finding nothing.
So I built CineMood — discover movies based on
your mood, not what's trending.

🔗 Live: https://cinemood-plum.vercel.app/

Stack

React + Node.js + Express
MongoDB
TMDB API
Vercel + Render

Biggest Challenge — CORS Error

Got this error connecting Vercel frontend
to Render backend:

blocked by CORS policy

Fix:

app.use(cors({
  origin: 'https://cinemood-plum.vercel.app'
}))

2 hours of debugging. One line fix. 😅

What I Learned

Set up CORS on day one, not after deployment
TMDB API needs caching for performance
Vercel + Render = best free fullstack combo

Found this helpful? Drop a ❤️