At OCP Group, I designed and built an end-to-end AI platform for managing more than 10,000 fixed assets. The work covered the existing-process diagnosis, data pipeline, predictive model, dashboards, and deployable web application.
I built an ETL pipeline to import SAP and Excel files into PostgreSQL, with data-quality checks, anomaly handling, traceable import sessions, and dataset versioning. For asset prioritization, the selected model reached a ROC-AUC of 0.975, and captured 99% of positive cases in the top 10% of highest-priority assets.
I added decision-support features including asset-level SHAP explanations, what-if simulations, and a Power BI cockpit. The platform also included a RAG assistant with source citations and an OCR workflow with human validation; the final decision remained with the user.
At ANZAR, I developed a web platform connecting travelers and worked on its relational and NoSQL data structure, front end, back end, and business-data APIs. My projects also explore online fault detection, digital twins, and real-time monitoring, alongside my Master's studies in Data Science for Smart Industry.

