Chahla Tarmoun

- Designed and implemented anomaly detection algorithms for trade surveillance, analyzing the impact of bank orders on market prices to identify and quantify potential price manipulation. - Built and optimized data pipelines processing millions of daily trade events, ensuring high data quality for model training and evaluation in a regulated banking environment. - Collaborated with compliance and quantitative teams to translate technical outputs into actionable insights for risk mitigation

- Developed a pipeline for extracting abstracts from scientific literature (PubMed) using APIs. - Conducted Named Entity Recognition (NER) to identify food and disease entities and linked them to biomedical databases (e.g., FooDB, MeSH). - Fine-tuned Transformer models (BERT, BioBERT, RoBERTa) for Relation Extraction (RE) tasks, classifying relationships such as 'cause' or 'treat.' - Improved the precision of personalized nutrition recommendations by automating the extraction of key relations between food and diseases. - Collaborated with domain experts for annotation and validation of key datasets, contributing to robust model training.

- Data collection via web scraping using Python, BeautifulSoup, and Selenium. - Market research on agricultural product pricing, including analysis of margin discrepancies among sellers. - Price prediction for wholesale markets using machine learning and deep learning models with Sklearn, Keras, and TensorFlow.