Filippo Leonelli
Machine Learning Engineer
Machine Learning Engineer
MazeGenerator.me is more than just a puzzle tool — it’s a playground for creativity, problem-solving, and fun. Whether you’re a teacher looking for engaging classroom activities, a parent wanting quick printable games for kids, or a puzzle enthusiast who loves the challenge of a good labyrinth, this generator makes it effortless to create unique mazes in seconds. What makes it even more exciting is that the entire project was vibe-coded with the help of AI tools — built with a mix of experimentation, intuition, and cutting-edge technology. It’s not just functional; it’s a testament to how modern AI can empower indie creators to build something fun, useful, and accessible to everyone.
Try it!
At the heart of this project lies a convolutional autoencoder, a sophisticated deep-learning model renowned for its denoising capability. Traditional methods, such as thresholding, often fall short when faced with the delicate task of preserving character separation in lighter fonts. These methods tend to merge characters or fail to detect them entirely, leading to significant information loss and errors in the digitized text.
cView, the micro-framework designed for crafting static sites and landing pages with ease. Optimized for projects that require a few pages and manageable content through a GraphQL-based Headless CMS, cView aims to simplify your web development process.
This repository contains a Python wrapper for running Simulink models from MATLAB within a Python environment. It enables setting parameters for a Simulink model, running simulations, and retrieving results for further analysis in Python. The wrapper uses MATLAB Engine API for Python to interact with MATLAB.
This repository provides a complete setup a for a semantic search engine for your PDFs. Semantic search leverages the potential of sentence embeddings derived from Large Language Models (LLMs) such as GPT-4 and BERT. This search approach centres on finding meaning matches within extensive text databases rather than focusing on specific keywords.
This repository contains a Python wrapper for running Simulink models from MATLAB within a Python environment. It enables setting parameters for a Simulink model, running simulations, and retrieving results for further analysis in Python. The wrapper uses MATLAB Engine API for Python to interact with MATLAB.
The Stereocamera Visual Odometry for Drone Localization Repository is a dedicated resource focused on implementing visual odometry techniques using a stereoscopic camera setup for precise drone speed estimation. Inside the repository, you will find feature extraction and matching algorithms, as well as calibration and triangulation methods.