Data Science

Data Science extracts actionable intelligence and predictive power from the university’s most complex data challenges. Blending advanced statistical modeling with the frontier of machine learning, this pillar specializes in customizing open-source models for highly specialized academic disciplines. Beyond predictive modeling, it provides the computational framework for decision science, allowing researchers to optimize complex operations and simulate outcomes with mathematical rigor.

• Decision Science & Operations Research: We deploy advanced mathematical optimization frameworks (Linear and Mixed-Integer Programming) alongside complex time-series forecasting to solve large-scale logistical, experimental, or resource allocation problems.

• Responsible AI & Explainability: We audit algorithmic bias, implement technical explainability metrics, monitor fairness, and author formal Model Cards to meet regulatory and institutional compliance standards.

• Advanced Data Lifecycle Engineering: We architect data processing and feature engineering pipelines to ingest, clean, and transform massive, heterogeneous datasets for downstream machine learning workloads.

• Model Evaluation & Adversarial Testing: We engineer rigorous evaluation harnesses, automated hallucination detection pipelines, and technical red-teaming protocols to stress-test model safety and accuracy.

• Open-Source LLM Fine-Tuning & Alignment: We provide deep-learning customization, specializing in parameter-efficient fine-tuning and advanced alignment training to tailor foundational models to specialized research domains.

For more information, please contact Caleb Reinking at creinkin@nd.edu.