Value Driven Data Science

Large language models excel at some tasks while struggling badly with other. At a time where it seems everyone is currently trying to make sense of AI, understanding and being able to explain exactly why this is the case, has become one of the most useful things data scientists can offer their stakeholders.

In this special reverse interview episode, Lauren Pearl switches roles with Dr Genevieve Hayes to explore why LLMs work the way they do and what that means for anyone using AI in their work.

You'll discover:
1. What your training data tells you about what AI can and cannot do [05:05]
2. Why high context problems are particularly hard for LLMs [06:28]
3. How AI gives you the right answer for the wrong reason [08:08]
4. Why some problems will always be beyond what LLMs can solve [14:22]

Guest Bio

Lauren Pearl is a business strategist, three-time founder and CFO advisor who helps start-up teams build data-driven businesses that thrive. She is also the resident start-up finance expert at NYU Stern’s Berkley Center for Entrepreneurship, where she teaches financial modelling to founders, and the co-host of The Growth-Minded CFO podcast.

Links

• Connect with Lauren on LinkedIn
• Subscribe to The Daily CFO newsletter
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Building something valuable with AI doesn't automatically make you valuable. If the value sits in your creation and your employer owns your creation, you're the one who's dispensable, not the thing you built. This is the AI chicken nugget problem and many data scientists who are impacted by it haven't even realised it yet.

In this episode, Lauren Pearl joins Dr Genevieve Hayes to explore where real expertise still lives in the age of AI and what data scientists should actually be focusing on to build a sustainable career.

You'll discover:

1. What the AI chicken nugget problem is and why it matters [02:31]
2. Why vibe coding is not a rare skill and never will be [05:12]
3. Why obsessing over something rare and valuable is your best career bet [17:17]
4. The skills worth investing in for a sustainable career [22:18]

Guest Bio

Lauren Pearl is a business strategist, three-time founder and CFO advisor who helps start-up teams build data-driven businesses that thrive. She is also the resident start-up finance expert at NYU Stern’s Berkley Center for Entrepreneurship, where she teaches financial modelling to founders, and the co-host of The Growth-Minded CFO podcast.

Links

• Connect with Lauren on LinkedIn
• Subscribe to The Daily CFO newsletter
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Machine learning is what most data scientists know best. But when it comes to building systems that enable better decision-making, machine learning is just one piece of the puzzle. And leaning on it too much can leave your entire decision system exposed.

In this Value Boost episode, Adam DeJans Jr joins Dr Genevieve Hayes to walk through the eight disciplines that make up the modern decision stack and where data scientists should focus to close the gaps in their own skill set.

You'll discover:

1. The eight disciplines that make up the modern decision stack [01:24]
2. The most overlooked layer in most decision systems [04:47]
3. Why optimization textbooks are teaching data scientists the wrong things [06:17]
4. What to learn next if you want to build better decision systems [09:45]

Guest Bio

Adam DeJans Jr. is a decision scientist and optimisation expert who has led high-stakes decision systems in complex, uncertain environments, at companies including Amazon, Toyota and Ford. He is the co-founder of AI and decision intelligence consultancy Bit Bros and co-author of The Decision Factory: A Novel About Decisions Under Uncertainty.

Links

• Connect with Adam on LinkedIn
• Bit Bros 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 are trained to build better models. But stakeholders don't wake up hoping for a better model. They wake up hoping to make better decisions. And the gap between those two things is where most data science value quietly disappears.

In this episode, Adam DeJans Jr joins Dr Genevieve Hayes to share what it actually looks like when an organisation stops optimising for better predictions and starts building systems that enable better decisions - that is, decision factories.

You'll discover:

1. What separates a decision factory from a machine learning pipeline [02:11]
2. Why uncertainty is an input into good decision making [10:51]
3. How to test a decision policy before it costs you real money [12:11]
4. Why giving stakeholders options beats giving them the optimal solution [18:48]

Guest Bio

Adam DeJans Jr. is a decision scientist and optimisation expert who has led high-stakes decision systems in complex, uncertain environments, at companies including Amazon, Toyota and Ford. He is the co-founder of AI and decision intelligence consultancy Bit Bros and co-author of The Decision Factory: A Novel About Decisions Under Uncertainty.

Links

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

Choosing which AI tools to buy is the easy part of an AI strategy. The hard part - and the part most organisations are ignoring - is everything that sits beneath the tools, from platform and inference decisions to hardware and sovereignty.

In this Value Boost episode, Victor Coimbra joins Dr Genevieve Hayes to share his framework for cutting through the noise of AI tool selection to the strategic decisions that will actually determine whether an organisation's AI future succeeds or fails.

You'll discover:

1. The four layers of an agentic AI strategy and why most organisations only think about one [01:53]
2. The three symptoms that signal an organisation's AI strategy is breaking down [05:52]
3. The four questions that reveal what an organisation's AI strategy is actually missing [08:21]
4. The one question data scientists should lead with when advising stakeholders on AI [10:43]

Guest Bio

Victor Coimbra is a Partner and CTO at Artefact, the world’s largest pure-play AI consulting firm and co-founded the firm’s Latin American operations. In 2024, he was recognised in the Forbes 30 Under 30 Brazil list for his outstanding contributions to AI innovation.

Links

• Connect with Victor on LinkedIn
• Victor's article on the four-layer framework
• Artefact website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

The debate about whether AI will replace human workers has already been settled - not by academics or futurists, but by the organisations that fired their humans, discovered AI couldn't do what they needed, and quietly hired them back. The future isn't AI replacing humans. It's humans and AI working together in ways that neither could manage alone.

In this episode, Victor Coimbra joins Dr Genevieve Hayes to share what hybrid agentic organisations actually look like in practice, and what data scientists need to do to position themselves at the centre of them.

You'll discover:

1. Why thinking of AI as a tool rather than a coworker is the mindset holding most organisations back [03:58]
2. The four archetypes that determine which tasks belong to humans and which to agents [08:17]
3. How AI is turning data scientists back into scientists [18:24]
4. The two skills that will define an indispensable data scientist in a hybrid organisation [27:01]

Guest Bio

Victor Coimbra is a Partner and CTO at Artefact, the world’s largest pure-play AI consulting firm and co-founded the firm’s Latin American operations. In 2024, he was recognised in the Forbes 30 Under 30 Brazil list for his outstanding contributions to AI innovation.

Links

• Connect with Victor on LinkedIn
• Artefact 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 hallucinations get all the attention. But hallucinations are relatively easy to catch because the output is obviously wrong. The failure mode that should worry data scientists more is when the agent uses facts that are true to draw conclusions that are false, producing outputs that look perfectly fine. This is known as silent correctness.

In this Value Boost episode, Jia Huang joins Dr Genevieve Hayes to explore why silent correctness is the most dangerous failure mode in agentic AI systems and what data scientists can do to catch it before it causes serious harm.

You'll discover:

1. Why silent correctness is harder to catch than a hallucination [04:17]
2. Why sampling and auditing are non-negotiable in agentic systems [07:09]
3. Four techniques data scientists can use to catch silent failures [09:28]
4. The one safeguard every agentic AI system should have [11:10]

Guest Bio

Jia Huang is a lead research engineer at A*STAR, Singapore's Agency for Science, Technology and Research, and is the author of multiple books on AI engineering and agent design, including Designing AI Agents and RAG from First Principles. His work focuses on turning agentic AI from impressive demos into reliable, auditable, and value-producing engineering systems.

Links

• Connect with Jia on LinkedIn
• Follow Jia on Substack
• Agent Design Pattern Society (ADPS) website
• Jia's AI agent design position paper
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

The shift to agentic AI doesn't make data science skills obsolete. But it does require data scientists to rewire how they think about familiar concepts, such as uncertainty, model evaluation and accountability, in their work.

In this episode, Jia Huang joins Dr Genevieve Hayes to explore what that rewiring actually looks like, and why data scientists are better placed than almost any other profession to make it.

You'll discover:

1. Why data scientists are better prepared for the agentic AI era than they might think [03:00]
2. How the data scientist's role is shifting from analyst to system designer [06:37]
3. The three types of uncertainty in agentic AI systems [11:58]
4. Why context engineering is the new feature engineering [23:19]

Guest Bio

Jia Huang is a lead research engineer at A*STAR, Singapore's Agency for Science, Technology and Research, and is the author of multiple books on AI engineering and agent design, including Designing AI Agents and RAG from First Principles. His work focuses on turning agentic AI from impressive demos into reliable, auditable, and value-producing engineering systems.

Links

• Connect with Jia on LinkedIn
• Follow Jia on Substack
• Agent Design Pattern Society (ADPS) website
• Jia's AI agent design position paper
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

A well-written blog post gets you noticed. But for data scientists who want to build authority that compounds over time, it's just the beginning. Every piece of writing is a potential stepping stone to something bigger - a conference talk, a book deal, or an opportunity you couldn't have anticipated.

In this Value Boost episode, Cynthia Dunlop joins Dr Genevieve Hayes to explore how data scientists can convert blog writing into bigger opportunities and what it actually takes to make the leap from blog post to book.

You'll discover:

1. How conference organisers actually find their speakers — and why blogging is the answer [02:35]
2. How acquisitions editors scout for authors and why you don't need a huge following [03:39]
3. The low risk way to find out if you're ready to write a book [06:57]
4. How each new opportunity compounds the authority you've already built [08:10]

Guest Bio

Cynthia Dunlop is the co-author of Writing for Developers and Senior Director of Content Strategy at ScyllaDB. She has co-authored four books for software developers and tech leaders and authored hundreds of articles for publications including TechCrunch, IEEE Computer, and The New Stack.

Links

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

For data scientists who want to build authority beyond their organisation, writing is one of the most powerful tools available. But in a world flooded with AI-generated content, simply publishing is no longer enough. The data scientists who stand out are the ones writing things no AI could have written.

In this episode, Cynthia Dunlop joins Dr Genevieve Hayes to share practical frameworks for writing blog posts that stand out, build genuine authority and actually get read.

You'll discover:

1. Why AI-generated content has made personal experience more valuable than ever [05:10]
2. The three Ps test for finding topics you can write about with genuine authority [08:56]
3. The blog post patterns that work best for demonstrating expertise [11:03]
4. How to use AI to improve your writing without letting it replace your voice [16:33]

Guest Bio

Cynthia Dunlop is the co-author of Writing for Developers and Senior Director of Content Strategy at ScyllaDB. She has co-authored four books for software developers and tech leaders and authored hundreds of articles for publications including TechCrunch, IEEE Computer, and The New Stack.

Links

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