Andre B.

The project aims to demonstrate responsible AI practices in healthcare-related machine learning, potentially improving model transparency and reducing bias in medical predictions.

In this project I am planning to apply the Microsoft Responsible AI toolkit to the scikit-learn Diabetes dataset, focusing on gaining practical experience with the toolkit and understanding its metrics in a real-world machine learning context. Key components to be explored include fairness assessment, interpretability techniques, and error analysis.

Originally developed in my free time to practice Python, Readability Optimizer is a prototype tool designed to help authors improve the readability of their texts.

This project marked one of my first endeavors in preparation for my upcoming Master’s in Applied Data Science (MADS) at the University of Michigan.

This prototype served as the foundation for an in-house solution, particularly for internal audit reports. This solution was expanded to import the audit report and analyze all observation paragraphs for readability.

The tool employs various readability tests and analyses to provide insights and suggestions for enhancing text clarity and comprehension, thereby improving written communication in professional settings.

Next step is to implement a local LLM to provide more text improvements recommendations.

Key findings indicate that TBIs, especially those involving loss of consciousness, alongside the presence of the APOE ε4 allele and male gender, significantly increase dementia risk, while higher education levels seem to offer some protective effects.

This project was designed to enhance our understanding of the outcomes following traumatic brain injuries, focusing on their determinants and consequences. We used causal inference to investigate the complex relationships between TBI and the subsequent risk of dementia, considering a range of demographic and clinical factors.

The complete report can be accessed here: Decode Dementia Report

This repository contains resources and examples for applying data analysis techniques in internal auditing. It aims to bridge the gap between data analysis theory and practical application in the field of internal audit including code examples, data sets and visualizations.

Our analysis of GDPR fines from 2018 to 2021 has revealed crucial insights for data protection practices, particularly in the healthcare sector.

Analyzing GDPR fines imposed by the European data protection authorities could reveal the main reasons and focus areas of the authorities for non-compliance and could allow our organization to timely address similar gaps in their data privacy strategy.

The complete presentation can be accessed here: Analyze GDPR Fines

The application of data analysis in internal audit has been a topic of increasing interest and discussion in my recent months. While numerous knowledge briefs and guides are available from respected organizations like the IIA and ISACA, these resources often present a more high-level and generic approach. As a professional with a background in data science, I’ve found that many existing resources in this area lack some depth and specificity to better bridge the gap between data analysis theory and its practical application in internal audit.

This observation has motivated me to create this comprehensive guide. This guide explores the application of data analysis techniques in internal auditing, adapting key concepts from academic research methodologies[5] to the practical world of internal audit focusing on two distinct concept.

In today’s world, algorithms are a big part of our daily lives. They suggest what we should watch and help decide if we can get a loan. These computer programs make many decisions that affect us all.

But what happens when these algorithms start to change the very world they are supposed to describe?

This is what we call algorithm ossification, and it’s important phenomenon to understand.

In the evolving field of artificial intelligence (AI) governance, two important concepts can get mixed up: AI System Classification and AI System Risk Assessment. This confusion can lead to problems in managing the related AI risks. This article aims to explain these concepts and how they relate to each other, using insights from major AI governance frameworks.

An overview of ChatGPT, delving into its underlying mechanisms, constraints, and practical applications.

This use case involved the development of a Data Ethics Checklist for data science projects, created as part of a university ethics course (MADS SIADS593). The checklist is designed to guide data scientists through ethical considerations at various stages of a project, from initial organization to final deployment. It incorporates key ethical principles such as transparency, accountability, fairness, and explainability, drawing from established frameworks and guidelines in the field of data ethics. The checklist is structured around the CRISP-DM model, providing specific, actionable questions at each project phase to ensure ethical practices are maintained throughout the data science workflow.

Originally developed in April 2021 as part of the University of Michigan’s Master of Applied Data Science program, SIADS501 - Being a Data Scientist course. This manifesto continues to guide my practice and is regularly updated to reflect new insights and evolving best practices in the field.

This document outlines my core principles, methodologies, and ethical commitments as a data scientist. It serves as a personal guide for navigating the complex landscape of data-driven decision making, ensuring that I consistently deliver high-quality, ethical, and impactful work.

This article addresses the frequently asked question “What is an (information security) risk?” from the perspective of an IS Auditor and Information Security Risk Manager. It presents three tailored explanations to accommodate varying levels of audience interest and time constraints.

The first is a concise formula, the second a comprehensive definition based on industry standards, and the third a detailed exploration using ISO 31000:2009 as a foundation. This multi-tiered approach offers flexibility in explaining information security risk across different professional contexts and audience needs.

This specialized certification demonstrates expertise in fraud detection, prevention, and investigation, including proficiency in financial transaction analysis, fraud law, and advanced investigative techniques. Skilled in implementing anti-fraud controls and managing internal investigations. The globally recognized CFE certification reflects a high level of fraud examination knowledge and adherence to ethical standards in the field.

Demonstrated proficiency in developing and implementing comprehensive data protection strategies, policies, and procedures. Adept at conducting thorough data protection impact assessments, efficiently managing data breaches, and ensuring organizational adherence to key regulations such as GDPR. Recognized for serving as a pivotal advisor on data protection matters and effectively liaising between the organization, data subjects, and supervisory authorities. Combines technical knowledge with strong communication skills to translate complex privacy requirements into actionable business practices, fostering a culture of data protection across the organization.

Advanced information security certified professional with comprehensive expertise across the Common Body of Knowledge (CBK) domains. Skilled in designing and managing cybersecurity programs, encompassing risk management, asset security, network protection, and security operations. Proficient in identity management, security assessment, and secure software development.