Ons Dekhil

•Developed a multi-agent AI accounting platform leveraging LangChain and LangGraph to orchestrate autonomous financial document processing, analysis, and decision-making workflows. •Built microservices and ETL pipelines with MLOps workflows to support model lifecycle management and intelligent decision-making. •Integrated prompt engineering experiments to refine LLM reasoning, optimize contextual understanding, and improve the accuracy of financial insights. •Implemented CI/CD pipelines to automate microservices deployment, enforce code reliability, and ensure seamless integration of AI components into the production environment. •Delivered dashboards and a mobile interface, unifying backend automation with real-timeinteraction. •Resulted in a scalable, production-ready AI system improving accuracy, reasoning depth, andoperational efficiency

-Monitored offshore aquaculture in Tunisia using Sentinel-1, Sentinel-2, and MODIS Aqua data, analyzing water quality and thermal trends. -Developed object-based algorithms in GEE and created spatial-temporal maps to evaluate aquaculture impacts.

-Developed and deployed predictive models for forecasting support cases. - Leveraged Azure ML, Power BI, and AutoML Studio for model development - Responsibilities included data preparation, model training, dashboard creation, and deployment