Orienta

YT demo link : https://youtu.be/VoWvZTIpxVc

Project Overview

Orienta is a platform designed to facilitate job and training search for people seeking new professional opportunities.
The platform also enables structured management and sharing of candidate profiles, optimized for recruiters.

The solution is composed of several core features:

1. CV Creation
2. Job Search
3. Training Search
4. Candidate List

---

Features

1. CV Creation

• Easily create your CV using predefined templates.
• Edit your CV directly on the platform.
• Download the CV in PDF format.
• Automatic saving: your CV is saved automatically, even if you leave the site before finishing.

2. Job Search

• Import your CV using simple drag & drop.
• AI analyzes your profile and automatically generates optimized queries to find jobs matching your skills.
• Results are retrieved via the Google Custom Search API.

3. Training Search

• Analysis of missing skills for a target position.
• AI suggests relevant training programs to fill these gaps.
• Training programs are automatically searched using optimized queries.

4. Recruiter

• Access to a marketplace of candidate profiles.
• Each profile is standardized and can be viewed by recruiters.
• Candidate CVs are available for consultation.

---

Architecture and Data Flow

CV Creation

1. CV editing through the frontend interface
2. Download in PDF format
3. Publication to the marketplace → Backend → Database

Job Search

1. CV import
2. CV analysis using Gemini
3. Generation of an optimized Google query via GPT-4o-mini
4. Job search through Google Custom Search API
5. Retrieval of matching job listings

graph LR
    %% Define styles
    classDef userNode fill:#0073b1,stroke:#fff,color:#fff,stroke-width:2px;
    classDef serverNode fill:#f5f5f5,stroke:#333,stroke-width:1.5px;
    classDef iaNode fill:#e8f4f8,stroke:#0073b1,stroke-width:1.5px;
    classDef apiNode fill:#f0f0f0,stroke:#34a853,stroke-width:1.5px;
    classDef resultNode fill:#fff,stroke:#0073b1,stroke-width:2px;
    
    %% Workflow elements
    A[👤 User] -->|Import/Drag & Drop CV| B[🌐 Frontend];
    B -->|Send CV| C[⚙️ Backend];
    C -->|Analyze| D[🤖 AI Model];
    D -->|1. Extract: experiences, target job| D;
    D -->|2. Optimized Google query| C;
    C -->|Search query| E[🔗 API<br/>Google Custom Search];
    E -->|JSON/XML Results| C;
    C -->|Job offers| B;
    B -->|📄 Display results| F[📋 Results];
    
    %% Apply styles
    class A userNode;
    class B,C serverNode;
    class D iaNode;
    class E apiNode;
    class F resultNode;
    
    %% Diagram style
    linkStyle default stroke:#0073b1,stroke-width:2px;

Loading

Training Search

1. CV import
2. Analysis of missing skills for the target position using Gemini and GPT
3. Generation of Google queries to find training programs matching missing skills
4. Search via Google Custom Search API

graph TD 
    %% Define styles with lighter grays but white text
    classDef userNode fill:#0073b1,stroke:#fff,color:#fff,stroke-width:2px;
    classDef frontendNode fill:#555555,stroke:#fff,color:#fff,stroke-width:1.5px;
    classDef backendNode fill:#666666,stroke:#fff,color:#fff,stroke-width:1.5px;
    classDef iaNode fill:#444444,stroke:#fff,color:#fff,stroke-width:1.5px;
    classDef apiNode fill:#4A4A4A,stroke:#fff,color:#fff,stroke-width:1.5px;
    classDef resultNode fill:#333333,stroke:#fff,color:#fff,stroke-width:2px;
    
    %% Main workflow
    A[👤 User] -->|1. Import CV| B[📱 Frontend];
    
    subgraph "User Interface"
        B -->|2. Send CV + Target Job| C[⚙️ Backend];
    end
    
    subgraph "Analysis & Comparison"
        C -->|3. CV Analysis| D[🤖 AI: Analyzer];
        D -->|4. Skill Extraction| D;
        D -->|5. Target Job Identification| E[🎯 Comparator];
        E -->|6. Required vs Current Skills| E;
        E -->|7. Gap Detection| F[📊 Query Generator];
    end
    
    subgraph "Training Search"
        F -->|8. Queries per Skill| G[🔍 Training API<br/>Google Custom Search];
        G -->|9. Raw Results| H[📦 Aggregator];
        H -->|10. Filtering & Ranking| H;
    end
    
    H -->|11. Recommended Trainings| C;
    C -->|12. Display Results| I[📚 Personalized Trainings];
    I -->|13. Consultation| A;
    
    %% Apply styles
    class A userNode;
    class B frontendNode;
    class C,H backendNode;
    class D,E,F iaNode;
    class G apiNode;
    class I resultNode;
    
    %% Global style for background
    style TD fill:#2D2D2D,color:#fff;
    
    %% Arrow styles
    linkStyle 0 stroke:#64B5F6,stroke-width:2px;
    linkStyle 1 stroke:#64B5F6,stroke-width:2px;
    linkStyle 2,3,4,5,6 stroke:#64B5F6,stroke-width:2px;
    linkStyle 7 stroke:#81C784,stroke-width:2px;
    linkStyle 8,9 stroke:#81C784,stroke-width:2px;
    linkStyle 10,11 stroke:#64B5F6,stroke-width:2px;
    linkStyle 12 stroke:#64B5F6,stroke-width:2px;

Loading

Candidate List

• Profiles are retrieved from the CV model stored in the database.

---

AI Models and Tools Used

• Gemini 2.5 Flash: CV data extraction
• GPT-4o-mini: generation of optimized Google search queries
• Google Custom Search API: retrieval of job and training results

---

Requirements and Installation

Requirements

• Docker (if using Docker Compose)
• Postgres 16.10
• Java 17
• Apache Maven 3.9.9
• Node.js 24.8.0
• npm

application.properties Configuration (read tools installation.md)

Once the API keys are created, add them to the backend configuration file:

backend/src/main/resources/application.properties

# API Keys for AI and Search
openai.api.key=YOUR_OPENAI_API_KEY
google.gemini.api.key=YOUR_GEMINI_API_KEY
google.api.key=YOUR_GOOGLE_API_KEY
google.search.engine.id=YOUR_SEARCH_ENGINE_ID

#### If using Docker 
docker compose up -d

#### Else 
Backend 
cd backend
mvn install
mvn spring-boot:run

Frontend 
cd cv-frontend
npm install
npm run dev