Siwar Jendoubi

– Mentored Doctor of Business Administration (DBA) candidates through the complete dissertation development process, providing guidance on research methodology, academic writing, and business plan formulation.

– Designed and implemented a Retrieval-Augmented Generation (RAG) application to automate the extraction and summarization of key clauses from complex financial contracts. – Deployed the solution as an API (FastAPI, vector database) with a React frontend, enabling business teams to query contracts in natural language. – Developed and implemented a new LLM guided clustering algorithm that uses insights and recommendations provided by an LLM. – Developed and implemented Natural Language Processing (NLP) and Deep Learning models for email classification, executing end-to-end projects from planning to package delivery with a success rate greater than 90% and achieving customer satisfaction. – Mentored and managed interns to analyze customer emails using active learning and generative AI, improve email preprocessing using NLP, and finetune deep learning models through transfer learning. – Mentored and managed a PhD student on a Generative AI project, he defended his thesis in January. – Developed and implemented a clustering solution for trading transactions data to help the client making the Murex trading system reconciliation more efficient, resulting in increased reconciliation efficiency by about 50%. – Planning project tasks, carrying out tasks, evaluating, testing, and validating solutions. – Write detailed reports about accomplishments, present solutions/results, and write CIR reports. – Write and publish scientific papers and participate to scientific conferences. – Animation of internal training about AI and machine learning.

– Build a multimodal tree species recognition approach from leaves and barks – Increase the recognition rate by about 20% compared to the existing solutions – Write and publish scientific papers and participate to scientific conferences.

– Build two influence maximization models to select a small number of influencers who can maximize the propagation of a given marketing campaign through Twitter. – Improve the quality of detected influencers compared to existing solutions. – Write and publish scientific papers and participate to scientific conferences.