Aicha ETTRIKI

Project: Creation of guidelines for model usage and implementation of an automated evaluation pipeline. Evaluation of
rules-based vs. LLM approaches on clinical data. Structuring and preparation of datasets. Contribution to writing a
scientific poster accepted for presentation at AACR 2026.

During my internship, I created a call bot that uses a MySQL database to provide responses. The call bot is designed to answer questions by querying the database and delivering the relevant information. I utilized the Tapas model from Hugging Face to enhance the bot's ability to understand and respond to natural language queries accurately. This project involved : -Using a pre-existing MySQL database and making some modifications, such as adding joins to improve query results. -Developing the logic for the call bot to interact with the database and retrieve information based on user input. -Integrating and fine-tuning the Tapas model for efficient and accurate interactions.

I had the opportunity to develop and deploy an image classification model. My main project was to determine whether an image had been retouched with Photoshop or not. To achieve this, I performed the following tasks: -Data Collection and Preparation: Research and selection of a dataset of images, cleaning and preprocessing the data. -Model Development (CNN): Using machine learning techniques to build a classification model. -Training and Validation: I optimized hyperparameters to improve accuracy and reduce errors. -Results Analysis (Accuracy and F1-Score).

-Responsible for member management. -Guided a team in exploring and deepening their knowledge in the field of computer science. -Responsible community manager.