Med Dhifallah

• Developing AI-powered automation and analytics solutions for startups and SMEs, focusing on NLP workflows, intelligent reporting, and data-driven applications.

• Building LLM-based systems and automation pipelines using Python, LangChain, Pandas, and machine learning frameworks to reduce manual analytical effort by up to 74%.

• • Deployed a Tableau dashboard used by 23 global teams, improving KPI tracking and reporting efficiency for the different Account Team.

• • Building an Automated EDA assistant with a Thinker–Worker architecture that applies rule-based heuristics for data profiling, cleaning, and fully logged transformations.

• • Built a resource-efficient AI prototype on Raspberry Pi and Arduino, enabling real-time embedded inference for a drone project and a line follower.

• • Improved model latency by deploying TensorFlow Lite + scikit-learn for RL-based tasks in constrained settings while documenting the process and serializing the data.