📝 AI Resume Creator

This is a full-stack web application developed as part of a university project for the course Fundamentals of Cloud Technologies. The application parses data from a given GitHub profile (with planned support for LinkedIn in the future) and uses an AI-powered system to generate a personalized PDF resume based on the provided information and the user's custom query.

🎯 Project Objective

As part of the assignment, our team was required to:

• Build a web application with both frontend and backend components;
• Deploy the application to the cloud using different providers for hosting and additional services;
• Integrate at least one external cloud service provided by a different vendor than the one hosting the app;
• Include AI or IoT functionality within the application;
• Use a cloud-based database for data storage;
• Document the architecture and technical choices made during the development;
• Collaborate effectively via GitHub, with clear task distribution and version control.

According to the requirements, the frontend and backend are hosted on different cloud environments. The AI module runs as a separate microservice deployed on a different provider, ensuring the usage of multiple cloud platforms.

🚀 Key Features

1. 🖥️ Frontend
   A user-friendly interface where users can:

   • Input their GitHub profile link
   • Write a custom query
   • Generate a personalized resume based on the provided data
   • View the resume in real-time
   • Download the final version as a PDF

2. 🔗 GitHub Integration
   A mechanism that allows users to input a GitHub link. The application then fetches relevant profile data — such as repositories, technologies used, and programming languages — via the GitHub API.

3. 🧾 Parsing Service
   Responsible for processing and transforming raw GitHub data into a structured format. This includes extracting and organizing details about the user's profile, projects, and skills for further use in resume generation.

4. 💬 AI Chat Service
   Leverages AI models to process both GitHub data and the user's custom text input. The service generates a LaTeX-formatted resume based on this information, ensuring the final document is tailored and professional.

5. 🔙 Backend API
   The server-side engine of the application. It handles:

   • Routing and logic
   • Data processing
   • LaTeX-to-PDF conversion
   • Communication between microservices (e.g., between the parsing service and the AI chat service)

6. 🗂️ Database
   Stores user-related data such as name, email, and password — with passwords securely stored using hashing to ensure privacy and security.

💻 Stack

🎥 Demo Video

demo-video.mp4

▶️ How to Use

1. Head to the home page — you will see the login/register menu.
2. Create an account, log in, or click "Don't Enter" to use the web-application without registration.
3. On the left side of the app, you'll find input fields:
   • Paste your GitHub or LinkedIn profile URL.
   • Optionally, write additional brief information about yourself.
     Recommended format:

     Name: John  
     Surname: Smith  
     Job: Backend Developer  
     University: TUKE  
     

4. Click the "Generate Resume" button to generate your resume. It will be displayed on the right side of the screen in an A4-style viewer. You can navigate through the pages using arrows.
5. To download your resume as a PDF, click "Download PDF". You can also click "..." to choose other available formats.
6. To clear all input fields, click the "Reset Parameters" button.
7. To log out, click the user icon at the top-left corner.
   • If you're logged in, select "Log out" from the dropdown — you'll be redirected to the login menu.
   • If you're using guest access, you'll see "Login" and "Register" buttons instead.

🐳 Docker Usage

1. Clone the repository to your local machine:

   git clone https://github.com/dmytro-varich/AI-Resume-Creator.git

2. Navigate to the project’s root directory:

   cd AI-Resume-Creator

3. Start the backend, parser service and frontend containers:

   ./start-app.sh

4. Verify these containers are running:

   docker container ls

   You should see output similar to this:

   CONTAINER ID   IMAGE                        COMMAND                  CREATED         STATUS         PORTS                              NAMES
   ef46f6e3ba13   ai-resume-creator-frontend   "/docker-entrypoint.…"   2 minutes ago   Up 2 minutes   0.0.0.0:3000->80/tcp               appfrontend
   cf4f4fe59df3   ai-resume-creator-backend    "dotnet ResumeCreato…"   2 minutes ago   Up 2 minutes   0.0.0.0:8080-8081->8080-8081/tcp   appbackend
   c4ab5cf8a2b8   ai-resume-creator-parser     "python ParsingServi…"   2 minutes ago   Up 2 minutes   0.0.0.0:5000->5000/tcp             appparser
   fc43bdb66f6e   postgres:15.11-alpine3.21    "docker-entrypoint.s…"   2 minutes ago   Up 2 minutes   5432/tcp                           postgres
   

5. Access the application in your web browser:

   • Frontend: http://localhost:3000.
   • Backend: http://localhost:8080.
   • Parser Service: http://localhost:5000.

6. To enter a running container:

   docker exec -it <container-name> /bin/sh

7. To stop and remove the containers:

   ./end-app.sh

🌐 Deployment

The system is deployed using a multi-cloud and hybrid infrastructure that distributes services across Microsoft Azure, Neon.tech, and a university-hosted server (TUKE). Each component is containerized with Docker, allowing consistent deployment and seamless cross-platform integration.

🏗️ Architecture Diagram

👥 Team

Name	Role	Description
Dmytro Varich	Frontend	Designed and developed the user interface and user experience of the application.
Ivan Tkachenko	Backend, DB	Responsible for server-side logic, inter-service communication and database management.
Nikita Pohorilyi	Parser	Specializes in extracting and processing data from various sources.
Arsenii Milenchuk	AI API	Focuses on development of a convenient AI Chat service.

🎨 Frontend

• 🎨 Design — the user interface was designed using Figma, while the logo and documentation visuals were created with Canva.
• ⚛️ React.js — served as the main framework for building the client-side of the application. External libraries like pdfjsLib were used to render PDF resumes directly on the website.
• 💨 Tailwind CSS — used for fast and flexible UI styling, enabling a responsive and modern minimalist design.
• 📦 Docker — used to containerize the frontend application, simplifying deployment and ensuring portability across environments.

---

🛠️ Backend & Database

• ⚙️ C# with ASP.NET Core — provides a robust, high-performance, object-oriented framework for building scalable and secure web APIs.
• 🧩 Entity Framework Core — used for ORM and database migrations; it integrates seamlessly with C# and supports code-first development.
• 🐘 PostgreSQL — chosen for its reliability and because it can be hosted for free on Neon.
• 📄 OpenAPI (Swagger) — enables automatic API documentation and testing tools.
• 🧾 Pdflatex — used as a LaTeX compiler to generate PDF files from LaTeX templates.

---

🕵️‍♂️ Parser

• 🐍 Python — chosen for its simplicity, rapid development capabilities, and rich ecosystem of libraries.
• 🧴 Flask — lightweight web framework used to build the REST API.
• 🍾 Flask-RESTX — used for route management and automatic API documentation.
• 📄 OpenAPI — integrated for API visualization and testing.
• 📦 Docker — used to containerize the parser service for portability and ease of deployment.
• ☁️ GitHub API — utilized to fetch user and repository data.
• 🔗 Requests — Python HTTP library used for making external API calls.

---

🧠 AI API

• ⚡ FastAPI — asynchronous Python framework with built-in support for automatic documentation via Swagger.
• 🧠 Ollama — used to interface with large language models; offers a wide range of models and flexible configuration options.
• 🌐 Nginx — acts as a reverse proxy and load balancer; selected for its speed, popularity, and seamless Docker integration.

This setup was deployed using docker-compose.
Nginx was particularly necessary because the GPU server was hosted within a university VPN.
An external server was configured to handle incoming requests and forward them to the internal GPU server, enabling remote access through the VPN network.