Achref Soua

• Developed a human gait data collection system using Arduino Nano 33 and Inertial Measurement Unit (IMU) sensors. • Enhanced data reliability through advanced filtering techniques and denoising autoencoders. • Experimented with deep learning architectures like GRU, LSTM, and Transformers to model sequential data, and optimized hyperparameters using tools like Optuna and Weights & Biases which led to increase models performance by 15%.

• Developed a meal recommender system utilizing Q-learning and evolutionary algorithms to generate personalized 7-day meal plans based on user data (age, sex, weight, height) and goals (weight loss, gain, or maintenance). • Incorporated a dynamic feedback system to adapt meal recommendations according to user preferences (e.g., vegetarian) and macronutritional goals, ensuring optimal calorie intake for weight management for a tailored experience.

- Scraped reviews and ratings from travel websites using Python libraries (BeautifulSoup, Selenium), adhering to legal and technical guidelines. - Organized and preprocessed large datasets by cleaning, normalizing, and structuring review text for sentiment analysis. - Trained sentiment analysis models (Logistic Regression, BERT) to classify reviews into positive, negative, or neutral categories, achieving high accuracy and F1-scores.