Aziz Taktak

At Bosch Rexroth, I developed end-to-end machine learning pipelines for delivery delay forecasting, from large-scale operational data through API-based deployment into internal business systems.

I've built automation for data cleaning, transformation, feature engineering, and data-quality validation, alongside monitoring systems designed to keep production models reliable. I work closely with engineering teams and business stakeholders to turn requirements into AI-driven solutions.

At ALECOL, I built a graph neural network for a children's learning platform, modeling student, teacher, and parent relationships to predict psychological evaluation outcomes. By incorporating parental background as graph nodes, I surfaced signals that student-only data could not reveal and translated results into dashboards for educators.

I also build practical LLM systems, including a RAG knowledge assistant with embeddings, vector search, document ingestion, and backend APIs, plus credit-risk explainability and behavioral analytics platforms.