Developed tools to extract information related to ESG topic within financial and extra-financial corporate reports.

Why?
• To compensate for the lack of quality and consistency in the non-financial information available (ESG indicators, asset-level information).
• To improve quantitative analyses on topics such as physical risk or impact assessment and/or portfolio alignment.

How?
• Refined corporate NACE classification using siamese transformers.
• Created a fine-tuned entity extraction tool to precisely extract relevant information related to assets, geolocations, metrics…within corporate reports.
• Leveraged few-shot learning techniques to detect incoherences with responsible investing guidelines.
• Publication of a Climate Finance Benchmark, assessing the relevance of RAG approaches to extract corporate climate related information.

Implemented a credit risk score using banking data for clients with limited to no credit history.

Why?
• Serve a larger number of customers.
• Capitalize on new available information (open banking) to assess creditworthiness and repayment behavior.

How?
• Analyzed the sociodemographic and risk profile of clients with no credit history.
• Developed a reusable and robust methodology for integrating banking data in Machine Learning models.

Leveraged state-of-the-art graph neural network approaches to improve suspicious activity detection.

Why?
• Support teams responsible for monitoring and combating money laundering and terrorist financing.
• Anticipate increasing regulatory pressure on these issues.

How?
• Benchmarked existing supervised, unsupervised and weakly supervised techniques to detected suspicious activities.
• Open-sourced ad-hoc study to assess the relevance of graph neural networks applied to incomplete transactional graphs.
• Improved existing detection tool by capitalizing on graph autoencoded embeddings with operational and conclusive results.