Yasmine Rmedi

■ Conducted exploratory data analysis (EDA) and feature engineering using Amazon SageMaker notebooks and processing jobs. ■ Built and optimised RAG pipelines using vector embeddings and Qdrant vector database for efficient knowledge retrieval. ■ Integrated LLMs (OpenAI, Mistral) into production NLP systems; designed and refined prompts to improve model relevance and output quality. ■ Implemented conversation memory and context management with Redis to enhance multi-turn AI interactions. ■ Automated web data collection and processing using Playwright for scalable data ingestion pipelines. ■ Containerised services with Docker; managed version control and collaboration via Git/GitHub.

■ Built an upsell predictive model with PySpark ML achieving 97% accuracy and 91% F1-score, deployed to production via FastAPI and React. ■ Produced analytical reports and dashboards in Power BI for performance monitoring and strategic decision-making.

■ Built an NLP-based similarity microservice comparing user queries against MongoDB-stored data, reducing external API calls and improving response efficiency. ■ Developed a fraud detection model and a customer satisfaction dashboard in Power BI. ■ Secured API access using OAuth authentication.

Developed an automated fiscal invoice declaration reporting tool in collaboration with business teams.