Niccolò Balestrieri
(AI Specialist & Engineer)
I'm just a nerd who love manga and AI.
Turning ideas into code.
(AI Specialist & Engineer)
I'm just a nerd who love manga and AI.
Turning ideas into code.
EunoAI is a modular, human-centered AI platform designed to promote digital well-being. Built entirely serverless on Cloudflare Workers & Pages, it integrates a custom Model Context Protocol (MCP) server enabling tool-based AI interactions such as screen-time reminders, digital detox tips, and mindfulness modules — without relying solely on the LLM.
Complete Retrieval-Augmented Generation system with streamlit web interface
This project focuses on the visual analysis of moving vehicles, specifically on model-based car localization and 3D road reconstruction using a single calibrated static camera
In this project, I present a detailed analysis of the ArXiv dataset, a community-supported resource founded in 1991 and operated by Cornell University. By leveraging cutting-edge technologies, the goal is to extract meaningful insights from the interconnected academic knowledge within ArXiv.
This project focuses on designing online learning algorithms to optimize a marketing campaign, including pricing strategies and bidding in auctions.
This project focuses on extracting geometric and metric information from an image depicting a cylindrical vault in Palazzo Te (Mantua), Italy. We use a variety of techniques, from feature extraction (edges, corners, lines, ellipses) to calibration (horizon line, vanishing points, camera matrix), concluding with a partial unrolling (unfolding) of the cylindrical surface.
This project explores the utilization of deep learning techniques in analyzing Photoplethysmography (PPG) signals to identify cardiovascular abnormalities, specifically Premature Atrial Complex (PAC) and Premature Ventricular Complex (PVC).
This repository contains the work focusing on the classification of leaves as healthy or unhealthy using deep learning techniques.
Hitori is a logic puzzle that originated in Japan, the puzzle involves eliminating numbers from a grid so that no number appears more than once in any row or column. At the beginning of the puzzle, some numbers are already filled in, and the solver must use logic and deduction to determine which numbers to eliminate. The solved puzzle will have all remaining numbers arranged so that there are no duplicates in any row or column.
Moon Patrol (ムーンパトロール, Mūn Patorōru) is a 1982 arcade video game developed and released by Irem.
The provided dataset consists of 48,000 time series, each 2,776 data points long due to padding. These series are divided into six categories (A-F), each representing a different domain. The actual length of each time series varies, reflecting the diverse nature of the dataset.
This project will explore the SQuAD 2.0 dataset and develop a question answering model using deep learning techniques; and loading the dataset and preprocessing the data, followed by training, evaluating and comparing our models.
A data visualization application using Dash! TropiViz is an application I've developed to simplify the analysis of time-series data from CSV files.
This project will present an analysis of an end-to-end project related to a dataset concerning information on drinking water potability.
The e-Mobility system reduces the environmental impact of transportation by enabling efficient electric vehicle charging services. Through standardized APIs, eMSPs, CPOs, and DSOs collaborate to manage charging availability, energy sources, and pricing.
The focus of the project, is to analyse the directed graph of the 2004 USA political elections. A directed graph is a graphical representation of a set of objects in which some pairs of the objects are connected by links that have a specific direction.
LinkedIn
Github
Instagram