Yassine GANA

During my final-year Master’s internship at the LIST3N Research Lab at the University of Technology of Troyes, I am contributing to a research project focused on real-time 3D reconstruction for autonomous drone-based victim detection in disaster environments. -Working on real-time 3D scene reconstruction for a drone-based system designed to assist in victim detection and localization in disaster environments. -Developing a reconstruction pipeline using Gaussian Splatting, aimed at achieving live, incremental updates instead of the traditional declarative approach. -Leveraging Structure from Motion (SfM) techniques and tools like COLMAP to extract camera poses and sparse geometry from aerial imagery. -Investigating integration of SLAM methods for continuous mapping from live drone feeds, with a focus on improving speed and robustness for real-time performance.

-Developed dynamic and responsive web pages using Next.js and React.js. -Designed and implemented FastAPI endpoints to support backend microservices in Python. -Integrated various third-party APIs, including Google Analytics, Stripe, and Gemini Chatbot. -Collaborated with a large team following Git best practices for version control and code management. -Executed basic cloud and DevOps tasks, such as writing Terraform code to provision Google Cloud Platform resources (e.g., Cloud Build and Cloud Run instances).

During my international internship at Sanskarm University in India, I had the opportunity to conduct an extensive study on applying feature extraction techniques and machine learning algorithms for corn leaf disease classification. -Developed a comprehensive study on applying feature extraction techniques (HOG, LBP, GLCM, GLRLM) with SVM and KNN algorithms for corn leaf disease classification. -Used and improved a balanced dataset of 3,399 corn leaf images, including healthy, blighted, and rusted samples. -Conducted extensive experiments to optimize parameters for various feature extraction techniques and machine learning algorithms. -Gained hands-on experience in image preprocessing, feature extraction, and machine learning model training and evaluation. -Created a software that does the classification of corn leaf images.

During this experience, I served as a backend developer for the muPeers online project, collaborating with the project owner (based in France) and another front-end developer. - Project focused on creating a digital platform dedicated to finance. - Primary responsibility involved developing data collection features from an external API (metaModeling) and storing data in the project's database. - Implemented a precise and optimized logic for data collection due to the substantial volume of data (approximately 40,000 companies). - Additionally, responsible for implementing a feature for regular data updates. - Designed storage APIs and display APIs for consumption in the front-end interface. The technologies that were used for the project : Python with FastAPI, Pandas, and MongoDB.

First concrete experience in the professional world as a Full Stack Web Developer at Next Consult Company. * Web-Based Inventory Management Application: - Independently created a web-based inventory management application. - Covered both front-end and back-end development. - Tasks included importing items via CSV files, conducting counting and verification, managing counting personnel, and presenting exports on inventory statuses. * Boat Travel Tickets Management Project: - Collaborated with a dynamic team on a project related to the management of boat travel tickets. - Role focused on the backend, involving: - Designing backend operations. - Handling the external GNV API and integrating it into the project. - Preparing other APIs for frontend calls. - Developing CMS features for administration. Utilized the MEVN (MongoDB, Express.js, Vue.js, Node.js) technology stack for both projects.

This internship was conducted at the headquarters of the Tunisian Bank (siège de la Société Tunisienne de Banque) in order to obtain my bachelor's degree.. The project took the form of a web application for managing client credits. Speaking about the technologies, the project was implemented using the MERN stack (MongoDB, ExpressJs, ReactJs, and NodeJs).