Behind the Scenes

How I Was Made

A look at the AI technology and architecture powering this Digital Twin CV. Built with modern web technologies and large language models.

The Digital Twin Concept

A Digital Twin CV is an AI-powered representation of a professional's career and expertise. Unlike a static resume, it can engage in natural conversations, answer specific questions, and provide contextual information about experience, skills, and projects. The AI is grounded in a structured knowledge base to ensure accuracy and relevance.

Technology Stack

Frontend
Modern React with type safety and utility-first styling
React 19TypeScriptTailwind CSS 4Woutershadcn/ui
Backend
Type-safe API layer with end-to-end TypeScript
Node.jsExpresstRPCDrizzle ORM
AI & LLM
Large language model with retrieval-augmented generation
GPT-4RAG ArchitecturePrompt Engineering
Database
Structured knowledge storage with semantic search
MySQL/TiDBKnowledge BaseVector Search

How the AI Works

1
Knowledge Base

Career information is structured into categorized entries covering experience, skills, education, and projects. Each entry includes metadata and keywords for efficient retrieval.

2
Query Processing

When you ask a question, the system searches the knowledge base using keyword matching and semantic relevance to find the most relevant information.

3
Context Assembly

Relevant knowledge entries are assembled into a context window, providing the AI with specific, accurate information about Leandro's background.

4
AI Response Generation

The LLM generates a response using the provided context, ensuring answers are grounded in real career data rather than hallucinated information.

Key Features

RAG Architecture
Retrieval-Augmented Generation ensures responses are based on actual career data, not AI imagination
Conversational Memory
The chat maintains context across messages for natural, flowing conversations
Source Attribution
Responses include references to the knowledge sources used for transparency
Real-time Processing
Fast response times through optimized query processing and caching

Built by Conversation

This Digital Twin CV was created through an iterative process of AI-assisted development. The combination of modern web frameworks, type-safe APIs, and large language models enables a new way to present professional credentials—one that's interactive, informative, and engaging.

React + TypeScriptGPT-4 PoweredRAG Architecture