Value Driven Data Science

These days, every organisation wants to describe themselves as "AI-first". But in the rush to find opportunities to use AI, it can be easy to forget that AI isn't always the right answer. 

In this Value Boost episode, Santosh Kaveti joins Dr Genevieve Hayes to explore the situations where AI isn't the answer, how to recognise them, and how to have the conversation with stakeholders who are convinced it is.

In this episode, you'll discover:

1. The types of problems where AI consistently falls short [01:36]
2. How to recognise when AI is the wrong tool for the job [04:46]
3. Why most AI conversations eventually lead back to data, people and processes [06:25]
4. How to push back on an AI solution without losing stakeholder confidence [09:43]

Guest Bio

Santosh Kaveti is the CEO and Founder of ProArch, a technology consultancy that helps enterprises operationalise AI securely and at scale. His expertise spans critical infrastructure industries, including power generation, manufacturing and healthcare, where he has seen firsthand how AI can drive business transformation in complex regulatory environments.

Links

• Connect with Santosh on LinkedIn
• ProArch website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Organisations today have no shortage of AI ideas. What they lack is the ability to turn those ideas into production-ready systems that deliver real business value.

For data scientists trying to get AI projects off the ground, understanding why that gap exists is as important as the technical work itself.

In this episode, Santosh Kaveti joins Dr Genevieve Hayes to share what organisations consistently get wrong when embarking on AI initiatives, and what data scientists can do to help get it right.

In this episode, you'll discover:

1. Why organisations with great AI ideas still fail to deploy them [02:16]
2. What history tells us about where the current AI wave is heading [09:48]
3. The real cost of bolting AI onto systems that weren't designed for it [13:42]
4. How to forge the cross-functional partnerships that get AI projects off the ground [22:21]

Guest Bio

Santosh Kaveti is the CEO and Founder of ProArch, a technology consultancy that helps enterprises operationalise AI securely and at scale. His expertise spans critical infrastructure industries, including power generation, manufacturing and healthcare, where he has seen firsthand how AI can drive business transformation in complex regulatory environments.

Links

• Connect with Santosh on LinkedIn
• ProArch website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

AI can get you to 60% of a finished output in minutes. But getting from 60% to 100% - the part where real insight lives - is where human expertise becomes the deciding factor. And the more expertise you bring, the further AI can take you.

In this Value Boost episode, Brent Dykes joins Dr Genevieve Hayes to apply his Four Zones of AI Productivity framework to the insight generation process and explore what it means for data professionals who want to position themselves as strategic advisors.

In this episode, you'll discover:

1. The Four Zones of AI Productivity and how they apply to insight generation [01:28]
2. Why AI can help you find an insight but can't generate an actionable one [06:39]
3. Why better AI tools will widen the gap between experts and novices [09:46]
4. How to use AI effectively in your insight generation process [11:44]

Guest Bio

Brent Dykes is the author of Effective Data Storytelling and the founder of AnalyticsHero. He has consulted with some of the world’s most recognised brands, including Microsoft, Sony, Nike and Amazon, and is a regular contributor to Forbes.

Links

• Connect with Brent on LinkedIn
• Effective Data Storytelling website
• Forbes article about the Four Zones of AI Productivity
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Most data scientists have been in this situation: you spend hours analysing a dataset, return to your stakeholder with your findings, and are met with a polite "that's interesting" - before your work disappears into a drawer, never to be seen again.

The problem usually isn't the analysis. It's that interesting observations and genuine insights are not the same thing.

In this episode, Brent Dykes joins Dr Genevieve Hayes to share the frameworks behind identifying and communicating insights that actually move organisations to act.

In this episode, you'll discover:

1. What makes an insight an insight and why only 5% of findings qualify [03:42]
2. The four dimensions that focus your analysis before you touch the data [11:25]
3. The six criteria for a truly actionable insight [15:06]
4. Why narrative outperforms an executive summary every time [19:14]

Guest Bio

Brent Dykes is the author of Effective Data Storytelling and the founder of AnalyticsHero. He has consulted with some of the world’s most recognised brands, including Microsoft, Sony, Nike and Amazon, and is a regular contributor to Forbes.

Links

• Connect with Brent on LinkedIn
• Effective Data Storytelling website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Building authority as a data professional doesn't require a large budget, a publisher, or even a large audience. But it does require a deliberate decision to share your thinking with the world and the patience to let that compound over time.

In this Value Boost episode, Prof. Rob Hyndman joins Dr. Genevieve Hayes to share how selectively giving away his work for free helped him become one of the most cited and influential statisticians in the world, and what data professionals at any stage of their career can learn from that approach.

In this episode, you'll discover:

1. Why Rob decided to give away his work for free from the start of his career [01:42]
2. How open source software multiplied the impact of his research [05:58]
3. Why authority building is a virtuous cycle and how to start it [09:47]
4. Why starting small is the right move [10:35]

Guest Bio

Prof. Rob Hyndman is one of the world’s most influential applied statisticians and a Professor in the Department of Econometrics and Business Statistics at Monash University. He has maintained an active statistical consulting practice for over 40 years, published over 200 research papers, co-authored more than 65 R packages and written five books on time series forecasting. He is also a Fellow of both the Australian Academy of Science and the Academy of Social Sciences in Australia.

Links

• Rob's website
• Otexts' website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Data scientists today are under pressure to adopt the latest tools - machine learning, LLMs, generative AI. But in the rush to embrace what's new, many are leaving some of the most powerful analytical tools sitting on the shelf. Tools that handle something modern AI largely can't: uncertainty.

In this episode, Prof. Rob Hyndman joins Dr. Genevieve Hayes to make the case for why rigorous statistical thinking remains indispensable in the age of AI, and what data scientists are giving up when they abandon it.

In this episode, you'll discover:

1. Why throwing data at an LLM is no substitute for building a model that understands the problem [04:27]
2. How combining classical statistics and machine learning can produce better forecasting results than either approach alone [08:22]
3. What data scientists lose when they stop thinking probabilistically - and why it matters for decision making [12:38]
4. Where to start if you want to strengthen your statistical foundations [25:10]

Guest Bio

Prof. Rob Hyndman is one of the world’s most influential applied statisticians and a Professor in the Department of Econometrics and Business Statistics at Monash University. He has maintained an active statistical consulting practice for over 40 years, published over 200 research papers, co-authored more than 65 R packages and written five books on time series forecasting. He is also a Fellow of both the Australian Academy of Science and the Academy of Social Sciences in Australia.

Links

• Rob's website
• Otexts' website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Over 100 episodes of conversations with world-class practitioners, a few ideas keep surfacing. Technical skill is necessary but never sufficient. The most valuable data professionals aren't the ones who build the best models - they're the ones who know which problems are worth solving. And the gap between those two things is where most data scientists are leaving value on the table.

In this milestone episode, Dr. Genevieve Hayes reflects on her career journey and the conversations that helped her arrive at these conclusions, with Matt O'Mara turning the tables to put her in the hot seat.

In this episode, you'll discover:

1. From statistician to machine learning advocate and back again - and what that journey revealed [09:49]
2. The crack in the data science skills market where significant value is hiding [18:59]
3. Why knowing which problems to solve matters more than knowing how to solve them [24:53]
4. The top three lessons from 100 conversations on what data science value actually means [33:49]

Guest Bio

Matt O'Mara is the Managing Director of information and insights company Analysis Paralysis and is the founder and Director of i3, which helps organisations use an information lens to realise significant value, increase productivity and achieve business outcomes. He is also an international speaker, facilitator and strategist and is the first and only New Zealander to attain Records and Information Management Practitioners Alliance (RIMPA) Global certified Fellow status.

Links

• Connect with Matt on LinkedIn
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Biased machine learning models don't just produce poor predictions. They can damage reputations, derail projects, and in high-stakes fields like healthcare, potentially cause real harm. Yet many data scientists don't check for bias until it's too late, missing the opportunity to address it at its source.

In this Value Boost episode, Serg Masis joins Dr. Genevieve Hayes to share practical techniques for detecting and mitigating bias in machine learning models before they become major problems for you and your stakeholders.

You'll discover:

1. The most common bias patterns to watch for [01:32]
2. How to diagnose whether bias exists in your model [04:44]
3. The three levels where bias can be addressed  [07:13]
4. Where to intervene for maximum impact [08:17]

Guest Bio

Serg Masis is the Principal AI Scientist at Syngenta, a leading agricultural company with a mission to improve global food security. He is also the author of Interpretable Machine Learning with Python and co-author of the upcoming DIY AI and Building Responsible AI with Python.

Links

• Serg's Website
• Connect with Serg on LinkedIn
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

When your machine learning model makes a decision that affects someone's medical treatment, financial security, or legal rights, "the algorithm said so" isn't good enough. Stakeholders need to understand why models make the decisions they do, and in high-stakes environments, model interpretability becomes the difference between AI adoption and AI rejection.

In this episode, Serg Masis joins Dr. Genevieve Hayes to share practical strategies for building interpretable machine learning models that earn stakeholder trust and accelerate AI adoption within your organisation.

You'll learn:

1. The crucial distinction between interpretable and explainable models [07:06]
2. Why feature engineering matters more than algorithm choice [14:56]
3. How to use models to improve your data quality [17:59]
4. The underrated technique that builds stakeholder trust  [21:20]

Guest Bio

Serg Masis is the Principal AI Scientist at Syngenta, a leading agricultural company with a mission to improve global food security. He is also the author of Interpretable Machine Learning with Python and co-author of the upcoming DIY AI and Building Responsible AI with Python.

Links

• Serg's Website
• Connect with Serg on LinkedIn
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Data scientists often jump straight to machine learning when tackling a new problem. But there's a foundational step that can dramatically increase your chances of project success and create more reliable business value. Mathematical modelling from first principles provides a low-cost scaffolding that can make your machine learning work more robust.

In this Value Boost episode, Dr. Tim Varelmann joins Dr. Genevieve Hayes to explain how building models from physics principles, like mass and energy conservation, creates a modular foundation that reduces computational costs and makes your work easier to understand.

In this episode, we explore:
1. What mathematical modelling from first principles actually means [01:20]
2. How to build modular models with different resolution levels [04:39]
3. When to add machine learning to first principles models [08:18]
4. The practical first step to incorporate this approach into your work [09:23]

Guest Bio

Dr Tim Varelmann is the founder of Bluebird Optimization and holds a PhD in Mathematical Optimisation. He is also the creator of Effortless Modeling in Python with GAMSPy, the world’s first GAMSPy course.

Links

• Bluebird Optimization Website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE