At SupplySustain, I build machine-learning and predictive-analysis workflows for complex numerical datasets, along with NLP pipelines that turn unstructured reports and technical documents into actionable information.
I also contribute to end-to-end AI pipelines with deep-learning and computer-vision components, combining textual and numerical data for comparative evaluation, risk assessment, and decision support.
At Pipekala, I developed analytical workflows for sales and customer data, including a recommendation-based lead-ranking workflow and a profitability-visualization framework. I also used RFM segmentation to identify high-value customer groups.
Through freelance projects, I delivered work across machine learning, scientific computing, numerical analysis, and electrical engineering. My scientific-computing platform project is modernizing MATLAB numerical code into a C#/.NET WPF application while preserving the original algorithms' numerical behavior.
