Data Leaders Who's Who: Marek Rucinski

---

Marek Rucinski

Deputy Commissioner Smarter Data, Australian Taxation Office (ATO)

In this rich article we are joined by ATO’s Deputy Commissioner Smarter Data, Marek Rucinski. In this role, Marek focuses on delivering value and industrialising data & analytics capabilities, of the 700 strong team and the whole organisation.

The ATO has established a careful approach to where and when AI projects should be implemented in order to deliver quantifiable value. Marek shares AI/ML use cases at the ATO and strategic advice on how to scale projects in order to avoid what he calls "proof of concept hell!" This rich interview also features insights on structuring teams, as well as really practical steps to make sure your AI ethics principles aren't just a poster on the wall, but actually implemented in day-to-day operations.

During this interview, Marek shares a wealth of information in answering the following questions:

• How do you determine which projects to use AI for and prioritise your projects?

• It is often difficult to quantify the ROI with AI and machine learning projects, is there anything you do specifically to measure the impact of an analytics product or project?

• Can you share any examples of AI/ ML solutions deployment or use cases where this has worked well?

• For large enterprises with legacy technology, how can we overcome the data hurdles in deploying AI?

• What are the biggest blockers to AI deployment?

• What are the limitations to scaling AI/ ML solutions in your opinion?

• In May 2021, only 22% of Data Futurology respondents polled in our Advancing AI series said they have automated monitoring of ML models in production. What would be your advice on why models need monitoring?

• How do you integrate AI ethics into the fabric of the organisation?

• What processes do you have in place to measure bias in decision making?

• What have you done at your company to build organisation-wide trust in your AI?

"INTEGRATION OF SMALL-SCALE PROOF OF CONCEPT TO AN INDUSTRIAL PROCESS IS A KEY POINT OF FAILURE FOR MANY ANALYTICAL PROCESSES."

"THESE MACHINE LEARNING MODELS IDENTIFY HIGH RISK GST REFUNDS AND RANK THEM BASED ON THE LIKELIHOOD AND CONSEQUENCE OF THE RISK."

• What is the best way to structure your data and analytics teams? What processes and methodologies are key to underpinning analytics project success?

• We are seeing a real demand across the industry for ML engineers, do you see that changing in the long term? Will ML just become a fundamental component to the data scientist's role?

• For those who are hiring on potential, what skillsets are most important to be able to build on?

• What would be your recommendations to leaders looking to attract diverse talent to their teams?

• Are there any data governance tools or frameworks you have been impressed with?

• What new technology and innovations do you see as being the most critical to your industry over the next 18 months?

• How do you find the balance between data governance and allowing analytics and data science teams the freedom to build models and provide insights to their business stakeholders?

• Finally, what work are you most proud of at the ATO?

To get the full PDF copy of this and every Data Leaders Who's Who interview coming out, join our Member Space. There you can also access past webinars in the webinar vault. If you are already a member, log in to access the PDF download.

---