Skip to content
Niš, Serbia

Resuma

Open-source desktop resume builder that writes and polishes your resume with on-device AI. Private, fast and fully offline on Windows, macOS and Linux.

Visit project Source
Role
Full-stack developer
Platform
Desktop (Cross-platform: Windows, Linux, macOS)
Categories
Desktop, Web, AI / ML
Year
2025
Resuma cover: title and project illustration on a dark background

01 The brief

Assignment

Build a privacy-first desktop app for creating resumes using local AI models and modern cross-platform frameworks.

Objective

Enable users to generate professional resumes instantly with on-device AI, ensuring full privacy, speed, and customization without reliance on cloud services.

Built with

  • Rust
  • TypeScript
  • React.js
  • Tauri
  • llama.cpp
  • TailwindCSS
  • Git
  • Vite

Project overview

Project Background

I created Resuma, an open-source desktop application for building professional resumes powered by local AI models. The challenge was to design a privacy-first tool that delivers AI-powered writing and formatting while keeping all data securely offline. This project combined AI integration, cross-platform development, and clean UI/UX to give job seekers a modern alternative to cloud-based resume builders.

Objective

  • Build a privacy-first resume builder powered by local AI models
  • Ensure all data stays on-device with no server dependencies
  • Provide customizable resume sections, layouts, and theming
  • Support dynamic PDF generation and preview with ATS-friendly formatting
  • Deliver a lightweight, cross-platform app for Windows, macOS, and Linux

Deliverables

  • Fully functional desktop app with AI-assisted resume writing
  • Local-only data storage and export (PDF, Markdown)
  • Advanced model configuration (context size, gpu_layers, mmap, etc.)
  • Modern, responsive UI with light/dark theming
  • Public GitHub repo, website, and installers for community use

Timeline and Milestones

  • Project Kickoff: Apr 22, 2025
  • Research: Apr 22 - Apr 26, 2025
  • Plugin Development (tauri-plugin-hwinfo): Apr 23 - Jun 21, 2025
  • Development Kickoff: Apr 26, 2025
  • AI Implementation: May 30 - Jun 7, 2025
  • Launch: Jul 1, 2025

Gallery - 02 images

Execution

Phase 1: Research & Planning

  • Defined target users: students, job seekers, and developers seeking privacy
  • Researched existing resume builders and their data-privacy shortcomings
  • Evaluated local AI frameworks (llama.cpp, GGUF models) for integration
  • Outlined app flow from onboarding → section editing → AI enhancement → PDF export
  • Skills Used: Market research, technical feasibility, architecture planning

Phase 2: Design & Development

  • Designed a clean, minimal interface in Figma with dark/light modes
  • Built the frontend in React + TailwindCSS with modular components
  • Developed the backend in Rust (Tauri) for cross-platform builds
  • Integrated llama.cpp with support for Qwen3 models (0.6B → 32B)
  • Implemented export features with LaTeX-inspired layouts for ATS compliance
  • Skills Used: UI/UX design, frontend + backend dev, AI integration, Rust tooling

Phase 3: Testing & Launch

  • Conducted cross-platform testing on Windows, macOS, and Linux
  • Optimized AI responses for clarity, trimming, and job description matching
  • Performed usability testing for first-time users (onboarding & workflow)
  • Packaged installers and published on GitHub + resuma.download
  • Released under GPLv3 and launched with community adoption and feedback
  • Skills Used: QA testing, deployment, open-source release management