A Comprehensive Approach to Quality in the Age of Data & AI | Testμ 2023

About the talk

Join us for a compelling session led by Bharath Hemachandran as he shares insights on 'Ensuring Quality in Data & AI: A Comprehensive Approach to Quality in the Age of Data & AI'. Discover how to maintain top-notch quality standards in the evolving landscape of data and AI. Don't miss out on this illuminating talk!

About Bharath Hemachandran

Bharath Kumar Hemachandran is a Quality Analyst and Principal Consultant at Thoughtworks. He has been working for over 16 years in the software industry in various roles, from that of a developer to the IT head of a real estate company. He loves to look at technology with a business mindset and solve real-world problems using technology. His passions include researching the use of Generative AI in Software Development and blogging.

Video Chapters

00:00 Introduction

45:19 QnA

49:16 Conclusion

Key Topics Covered

Introduction to Quality Assurance in AI and Data: The session begins with an introduction to the importance of adapting and extending traditional quality assurance methods to address the challenges presented by sophisticated AI systems.

Five Pillar Approach to Defining Quality: A framework for defining and measuring quality in AI and data ecosystems, which includes data quality, algorithm/model quality, infrastructure quality, ethical considerations, and compliance with regulations.

Challenges in Measuring Quality: The difficulty in defining and measuring quality in data and AI systems is discussed, with emphasis on the need for clear criteria beyond functional and cross-functional requirements.

Case Studies and Examples: Several case studies are presented to illustrate common pitfalls and challenges, including Google Flu Trends, the Apple Card gender bias issue, Equifax data breach, Cambridge Analytica scandal, and GDPR compliance.

Detailed Discussion on Each Pillar: The video delves into the importance of data profiling, cleansing, validation, and lineage tracking for data quality; the need for explainable AI, dealing with algorithmic bias, and model validation for model quality; best practices for infrastructure testing, disaster recovery planning, and security for infrastructure quality; the importance of ethical guidelines and guardrails in AI system development for ethical considerations; and ensuring compliance with laws and regulations, and the importance of data governance practices for regulatory compliance and data governance.

Impact on Traditional Roles: How the world of data and AI impacts the roles of developers, QA professionals, business analysts, and other stakeholders in the software development process.

Future Directions and Challenges: Discussion on the rapidly changing landscape of AI and data, the need for ongoing adaptation, and the importance of a holistic approach to quality.

Q&A Session: The video concludes with a Q&A session where the presenter addresses specific questions from the audience related to AI testing, model quality, and the application of traditional software quality methodologies to AI systems.