N/A

Predicting used car prices in the German market, creating a platform, and integrating a chatbot onto the site, all developed in German.

Development of a web application to analyze tweets and classify their sentiment as positive, negative, or neutral using natural language processing.

Creating a web platform for cultural center administrators to track visitors in various rooms, building profiles that include visitor information, notably their fingerprint, which will be recognized through a deep learning model. + Developing a web application to identify and interpret emotional expressions on visitors' faces, capturing their reactions to gather feedback.

Analyzing road accidents by leveraging SQL queries for data extraction and employing data cleaning and visualization techniques, including the use of Power BI. The findings are presented through interactive dashboards for comprehensive insights.

This project aims to implement an emotion detection system that utilizes computer vision and deep learning techniques within a Flask application that allows users to upload images and receive sentiment analysis results.

This project focuses on predicting promotions for subscribers using machine learning. It involves data preprocessing, analysis, and the development of a Streamlit app.

This project uses hyperspectral camera wavelengths (400nm to 1064nm) and machine learning with GANs in TensorFlow to detect anomalies and identify fraudulent honey by recognizing spectral patterns associated with adulteration.

This project focuses on diagnosing organizational challenges through the Kübler-Ross change curve, establishing support mechanisms and managing communication and leadership strategies. It also involves overseeing the implementation (pilotage) to facilitate a smooth transition during the change process.

This project focuses on Novel Class Discovery (NCD). It examines methods that transfer this knowledge either in one or two steps, employing techniques such as pseudo-labeling, self-supervised learning and contrastive learning. It also explores how NCD relates to fields like clustering and transfer learning to address key challenges in open-world learning and its applications in domains like cybersecurity, bioinformatics and anomaly detection.

This project processes a large-scale Yelp reviews dataset using PySpark for efficient storage, transformation, and analysis. NLP techniques (tokenization, stopword removal, TF-IDF) are applied for sentiment classification with Naïve Bayes, Linear SVC, and Logistic Regression. Folium, Matplotlib, and Seaborn are also used to explore user behavior and business distribution.

This project, developed during a week-long hackathon, focused on optimizing reinforcement learning algorithms for autonomous drone navigation in a dynamic grid environment. It implemented and compared PPO, Q-learning DQN and D* to enhance decision-making and obstacle avoidance, with a custom reward function improving learning efficiency and agent performance.