Walter Diele

Walter has been in data analytics for over 25 years. Starting off with hands-on traditional statistical analysis and BI, to building an AI data platform, to heading analytics departments and consultancy in AI in recent years. He has a strong focus on vision and people management, knowing that data scientist need enough freedom to be creative on one hand, and create business value and actually deliver on the other hand.

Next to this Walter can explain technical complex content to stakeholders in a way that they understand what is real and what is science fiction, something which is essential in expectation management. Key is to know what can add value today, where to invest now and what is better left for others to do.

He worked for a software company (SAS Institute, ’96-‘00), a data company (Acxiom, ’00-‘05) and a consulting company (Deloitte,’05-‘18), which gives a good understanding of the different perspectives of how to make AI/analytics work for real life business issues. The last two and a half years Walter uses this experience as a freelance AI consultant to help the Rabobank make the digital transformation.

Leading teams of data scientist and engineers is only successful with (at least conceptual) knowledge of the main tools and architectures. Since tooling develops quickly, continues innovation and learning is key. There is no such thing as efficient innovation, and new methods can be worthwhile sharing and can have business value before they are perfect in the eyes of the data scientist.

With tooling getting more powerful and user-friendly, the risk profile of advanced analytics shifts to the methodological questions, data quality and compliance. Especially -but certainly not limited to- black box models like CNN’s, it is of vital importance to be sure that no unwanted effects are in the training data and models are valid in their application.

AI, and advanced analytics in general, is too impactful to treat it as a topic for techies. The business, social and legal implications need to be taken into account when developing models for business decisions. For data scientists, topics like auditability, accountability and discrimination are typically not top of mind.

Walter has a master in Methodology and has a sharp view on the internal and external validity of (ML) models. He developed and facilitated AI-for-dummies courses for Deloitte professionals (including Deloitte Legal global partners) to share and broaden the understanding of AI so more professions can join the discussion on the safe use of AI. He developed a workshop on Bias in AI, helped the Rabobank develop their responsible AI framework and was speaker in the Rode Hoed’s debate on discriminating algorithms.