Xiao Ma

Context: HAIKU is a LegalTech start-up specializing in the development of a SaaS solution integrating AI to optimize legal research and analysis. Missions : Migration of the HAIKU application from Vue.js to React to improve performance and scalability. Optimization of the user interface for a better user experience. Tasks : - Implementation of new React components (dashboard, login and registration pages). - Use of hooks and management of API requests. - Compatibility testing on different browsers and devices. - Optimization of UI, functionality and bug fixes.

Context: As part of my final internship for the master’s degree in Computational Linguistics, I worked on the Nominalp project, which aimed to automatically identify oronyms in texts exploring the Alps. Missions: - Development of a named entity recognition model based on BERT and DistilBERT for oronym identification. - Optimization of pre-trained models to improve detection accuracy. Tasks: - Preprocessing of corpus data. - Manual annotation of oronyms. - Training and evaluation of named entity recognition models. - Performance comparison between different models, such as BERT, DistilBERT, and spaCy. Results: Significant improvement in oronym detection accuracy through fine-tuning of Transformer-based models.

The internship consisted in annotating in a structured way some Chinese articles from the journal "Economic Science" of Peking University using the XML language