I’m a developer, computer scientist, and machine learning engineer. As best I can, I want to use my skills to build a better world.
This is a page for quick access to my projects, you can have a look around if you’d like 😄
Web platform for managing multi-currency cryptocurrency portfolios. It allows to monitor mining resources and sale markets.
The full stack was developed and implemented by myself: the Website is built in jQuery and PHP, whatchdogs in C++ then self hosted on a linux server. Born from personal needs the project has been expanded (for fun) to become a standalone platform for public use.
Web Development
Reinforcement Learning
1
Yet another operating system. Implementation of an operating microkernel for ARM7 architecture systems. LL interaction with basic computer components with
interrupt handling and process scheduling. to make it more crunchy all functions are implemented with recursive approach.
Operative Systems
Reinforcement Learning
2
My first approach to computer vision. Revision and improvement of an open source program for the visual recognition of geometric shapes applied to puzzles. It allows also the reconstruction of the original image by composition of pieces.
Computer Vision
Reinforcement Learning
3
Starting from a video as input, the application returns the static background of the scene handling dynamism of the
scene in time. The algorithm works on-the-fly without assumptions or preprocessing of the frames, both with static and moving camera videos.
Computer Vision
Reinforcement Learning
4
Team participation in the AICrowd Flatland challenge 2021. The goal was to deal with a problem of trains coordination and navigation in
a complex railway environment. A multi-agent reinforcement learning approach was adopted in our implementation.
Reinforcement Learning
Reinforcement Learning
5
Custom implementation of "Context-Aware Retrieval-based Deep Commit Message Generation" paper for automatic commit messages generation. The implementation mix Neural Machine Translation with a Retrieval based approach using Transformers modules for the encoder-decoder architecture. The work include also an analysis of the result and a survey about the task.
Natural Language Proce
Reinforcement Learning
6
Game environment parameterised by a difficulty value to show the possibilities which an autonomous trainer agent can provide to a game developer in terms of automated processes during game development and enhancement of autonomous player agents.
Reinforcement Learning
Reinforcement Learning
7
Exploitation of the intrinsic information of a view to solve the SIDE (Single Image Depth Estimation) task and the NVS (Novel View Synthesis) task, achieving excellent results in both synthetic and real-world scenarios. The application finds its scope in enabling greater understanding of the world with limited sensory capabilities (e.g., in robots) and in enhancing visual contents.
Machine Vssion
Reinforcement Learning
8
Oct 2023 - Research Fellow in CVLab @University of Bologna
May 2023 - Intern researcher in LEAR (Learning Agents and Robots) Team @CYENS (Nicosia, CY)
Mar 2023 - Ended my MSc in Artificial Intelligence @University of Bologna
Oct 2022 - Thesis preparation abroad in Machine Intelligence Lab @UCL (London, UK)
Oct 2020 - Started Master degree in Artificial Intelligence @University of Bologna
