CloudHabits — Project Overview

CloudHabits started as a personal project with a simple goal: to help me learn and get hands-on experience with AWS. In the beginning, I set up my own cloud environment, configuring a server to run PHP, managing the database, and handling the deployment process from scratch. It was a great way to understand how web applications work beyond just writing code.

After a few months hosted on AWS, I decided to migrate the project to a web hosting provider so I could focus more on building features and improving the product itself.

Cloud Habits login/create account screen

The entire project was designed and developed by me — from the layout and user experience to both the frontend and backend implementation. It includes features such as user authentication, login flows, form validations, database integration, and other core web application functionality.

I also built it using StarshipCode, the IDE we use internally at Southern Cross Aviation, which is also available for external use. starshipcode.com (SSC).

CloudHabits became more than just a learning exercise — it turned into a complete project where I could combine infrastructure, design, and software development in one place.

Cloud Habits home screen

Daily Quotes App (iOS) — Project Overview

The app was developed to explore Apple's development environment and Swift programming language, tested on simulator and personal iPhone as an experimental project.

AI Agent (RAG) — Project Overview

RAG application developed with Python, LangChain, ChromaDB and Google Gemini. The system processes external PDF documents, splits them into semantic chunks, generates embeddings using the sentence-transformers/all-MiniLM-L6-v2 model, and stores them in a vector database. When a question is submitted, the most relevant document passages are retrieved through similarity search and provided as context to Gemini, enabling answers grounded in the source material. The project was developed and tested locally through command-line interface. The screenshot shows a test using the novel The Strange Case of Dr. Jekyll and Mr. Hyde by Robert Louis Stevenson as the knowledge source. The conversation is in Portuguese because that's how it was defined in the prompt.

Print screen of RAG pdf working.

As a security optional extension to the RAG project, Microsoft Presidio was integrated to detect and anonymize sensitive information before being indexed into the vector database, preventing personal data from being exposed to the retrieval system or the LLM. A sample PDF containing personal data was used for testing, and confidential fields were automatically replaced with anonymized placeholders. The assistant can still answer questions about the document's general content, but it cannot reveal anonymized information because those details are no longer available in the knowledge base.

Print screen of original document containing sensitive information.

Print screen of the same document after anonymization with Microsoft Presidio.