Skip to content
All work
Live · in developmentCase study

DevPilot AI

A full-stack SaaS product that helps developers get job-ready. It scores resumes against ATS criteria, reviews GitHub projects, runs voice-enabled mock interviews and matches candidates to jobs.

Live Demo: DevPilot AI (opens in a new tab)GitHub: DevPilot AI (opens in a new tab)
Role
Creator and lead developer: product, frontend, backend, AI integration
When
2026
Status
Live, with the frontend on Vercel and the API on Render. Built from March to June 2026 and still in development.
DevPilot AI landing page with an AI chat preview and a Start building button

Context

The problem

Developers preparing for jobs juggle separate tools for resume checks, portfolio feedback, interview practice and job search, and rarely learn why an application failed.

Solution

What I built

I designed and built the product end to end: a Next.js 16 app with ten dashboard modules, an Express and TypeScript API over MongoDB, real-time interview sessions over Socket.IO, and AI features on Groq's Llama 3.3 70B.

  • Resume ATS analyzer
  • GitHub project analyzer
  • Voice-enabled interview simulator
  • Job matcher
  • AI career mentor chat
  • Job Ready Score
  • Career roadmap generator
  • Portfolio builder
  • Cover letter generator
  • AI coding tools: code generator, bug finder, SQL assistant, documentation writer

Architecture

How it fits together

  1. Client
    Next.js 16 · React 19
    10 dashboard modules
    Socket.IO client
    live interview sessions
  2. API
    Express + TypeScript
    13 route groups
    Auth & guards
    JWT · GitHub OAuth · Zod · rate limits
  3. Services
    AI service
    Groq · Llama 3.3 70B
    GitHub & resume parsing
    REST API · pdf-parse
  4. Data
    MongoDB
    Mongoose · 6 models
    Docker Compose
    API + database
How a request moves through DevPilot.

Engineering

Key engineering work

  • Layered Express and TypeScript backend (routes, controllers, services, repositories) with 13 API route groups and Zod request validation.
  • Authentication with JWT and GitHub OAuth through Passport; Helmet, CORS and rate limiting on the API.
  • Resume pipeline: PDF upload with Multer, text extraction with pdf-parse, then AI analysis for an ATS score and specific fixes.
  • Live mock interviews: Socket.IO sessions carry questions and answers between the browser and the AI service, with voice input and output in the browser.
  • GitHub analyzer that pulls repository metadata, README and language data from the GitHub REST API for AI code review.
  • Weighted Job Ready Score that combines resume, GitHub and skills results into one number with a breakdown.
  • Swagger/OpenAPI documentation, Winston logging, and Docker Compose for the API and MongoDB.

Technology

Stack

  • Next.js 16
  • React 19
  • TypeScript
  • Tailwind CSS
  • Node.js
  • Express
  • MongoDB
  • Mongoose
  • Socket.IO
  • Groq · Llama 3.3 70B
  • JWT
  • GitHub OAuth
  • Zod
  • Docker

Status

Where it stands

Live, with the frontend on Vercel and the API on Render. Built from March to June 2026 and still in development.

Live Demo: DevPilot AI (opens in a new tab)GitHub: DevPilot AI (opens in a new tab)

Gallery

Screens and visuals

DevPilot AI sign-in page with GitHub sign-in and email login
Sign-in with email or GitHub OAuth
Illustration of an AI resume analysis report
Resume analyzer (illustration from DevPilot's landing page)
Illustration of an AI reviewing a code repository
GitHub analyzer (illustration from DevPilot's landing page)
Illustration of a studio microphone for voice interviews
Voice interview simulator (illustration from DevPilot's landing page)