Data Matas

What if scaling your data platform meant starting from scratch?

In this episode of Data Matas, John Napoleon-Kuofie, Analytics Engineer at Monzo, shares what it really takes to rebuild trust in your models - inside one of the UK’s fastest-growing digital banks.

From inheriting 1,000+ undocumented DBT models to challenging the purpose of legacy tests, John walks through the decisions his team is making to improve data quality, reduce noise, and future-proof their platform. It’s a candid, practical conversation about choosing clarity over complexity, and learning to say no to work that doesn’t deliver value.

👉 In this episode, you’ll learn how to:

Rebuild inherited models by starting with real-world concepts, not assumptions

Reduce alert fatigue by testing only what matters (and ignoring the rest)

Prepare your data architecture before layering on AI or self-serve tools

Design systems that future engineers can actually understand

Foster a culture where bottom-up innovation drives real change

Watch the full episode on YouTube, or listen on Spotify and Apple Podcasts.

👤 John Napoleon-Kuofie on LinkedIn
👤 Aaron Phethean on LinkedIn
🎧 Data Matas Podcast
🌐 Matatika Website

How do analytics engineers grow from writing SQL to designing entire data systems? And why do most companies still confuse tool mastery with real engineering skill?

In this episode of Data Matas, Oleg Agapov (Senior Analytics Engineer at Hiive) shares what it actually takes to go from junior to senior in data—beyond bootcamps, tools, and job titles. He breaks down the core shifts that matter most: thinking in systems, mastering data modelling, and creating structure that scales.

You’ll hear how Oleg is helping build self-serve analytics inside a fast-moving fintech startup, why most data work fails without discovery, and how AI is changing the role—but not replacing the role—of the analytics engineer.

🎙 Guest: Oleg Agapov, Senior Analytics Engineer at Hiive
Oleg has spent over 15 years in data roles, moving from analyst to engineer to analytics architect. Now at Hiive—a marketplace for private stock—he’s helping design scalable data models and BI tooling that enable business teams to self-serve. Oleg also mentors junior engineers and shares career guidance on LinkedIn weekly, offering a rare combination of technical depth and practical coaching.

⏱ Episode Takeaways & Timestamps

03:40 – Why analysts become engineers (and what tools don’t teach you)
Why Oleg moved from analytics into engineering, and how messy data triggered a career pivot.

08:15 – What junior vs senior actually looks like in analytics engineering
From DBT basics to architecture thinking—how your role shifts as you grow.

12:30 – Data modelling isn’t a feature, it’s a discipline
Why writing queries isn’t enough—and why most engineers only realise this at scale.

17:45 – Building analytics in a three-sided marketplace startup
How Oleg is helping Hiive build self-serve data for a unique financial model.

24:00 – How AI fits into the modern data workflow (and where it fails)
Why LLMs are better reviewers than creators—and why trust still starts with humans.

28:40 – The hidden risk of AI assistants in BI tools
What happened when an AI assistant hallucinated a metric—and nearly caused a decision error.

Who Should Listen?
If you’re an analytics engineer, data modeller, or anyone growing a data team inside a startup or scale-up, this episode will help you move beyond dashboards and into strategic, scalable thinking. Especially valuable for those navigating the shift from IC to senior roles.

📢 Like this episode?
Subscribe to the Data Matas YouTube channel for weekly insights from real data leaders.
Hit the bell to get notified when new episodes go live.
💬 What’s one skill you think separates senior engineers from juniors? Let us know in the comments.

🔗 Links & Resources

👤 Oleg Agapov on LinkedIn: https://www.linkedin.com/in/oleg-agapov
👤 Aaron Phethean on LinkedIn: https://www.linkedin.com/in/aaron-phethean/
🌐 Matatika Website: https://www.matatika.com
🎧 Data Matas Podcast: https://www.matatika.com/podcasts
📺 Data Matas YouTube Channel: https://www.youtube.com/@matatika/podcasts

What if your data team’s biggest risk isn’t technical debt - but mental exhaustion?

In this brutally honest conversation, Nik Walker (Co-op) shares how a culture of low-value work and constant reactivity burns out even the best data teams, and what to do about it.

About this episode
Data engineers aren’t struggling because the tech is hard—they’re struggling because the work often isn’t worth doing. In this episode, Nik unpacks why discovery matters more than dashboards, how to protect your team from the myth of speed, and why AI means nothing if no one trusts the data.

You’ll get practical, real-world insights from someone scaling data infrastructure in one of the UK’s most complex legacy organisations—without breaking the team or the bank.

About the guest
Nik Walker is Head of Data Engineering at Co-op, leading data transformation across a massive enterprise with deep legacy tech and community-first values. Known for his human-centric leadership style and vocal advocacy for neurodiversity in data teams, Nik brings humour, candour, and serious experience to the conversation.
🔗 Nik on LinkedIn

Timestamps and Key Learnings
06:15 – How Co-op builds safe, structured teams that don’t burn out
→ Create psychological safety with process, not platitudes

09:27 – Why neurodiversity awareness isn’t optional in data teams
→ 60% of data professionals are neurodivergent—your leadership style should reflect that

18:26 – What it really takes to trust your AI outputs
→ If the maths is off or the data’s wrong, no one will use your model

20:28 – Stop syncing everything in real-time
→ You don’t need real-time pipelines—you need right-time pipelines

28:07 – Discovery over delivery: how to stop wasting time and money
→ Methodical work delivers more value than rushed builds

Why listen
If you’re a data leader tired of firefighting, low-value tasks, or untrusted dashboards, this episode is for you. Nik offers tangible advice on building better systems, defending your team’s time, and navigating real-world transformation without breaking your people.

Subscribe and join the conversation
Like what you heard? Hit subscribe and tap the bell so you don’t miss future episodes.

Have you faced burnout in your data team? Drop a comment below and share your experience, we’d love to hear from you.

Links and Resources
🔗 Nik Walker on LinkedIn: https://www.linkedin.com/in/nikolaswalker/
🔗 Aaron Phethean on LinkedIn: https://www.linkedin.com/in/aaron-phethean/
🔗 Matatika: https://www.matatika.com
🎙️ Data Matas: https://www.matatika.com/podcasts/
📺 Data Matas YouTube channel: https://www.youtube.com/@matatika

In this episode, we reveal how IRIS Software is borrowing principles from quantum computing to build data systems that are adaptable, explainable, and innovation-friendly—without compromising business-as-usual.

This conversation with David Draper offers a rare inside look at how a major enterprise balances legacy systems with forward-thinking experimentation.

You’ll learn how modular architecture, protected innovation time, and explainable AI are reshaping the way IRIS builds its data infrastructure. If you’re a data leader trying to scale without spiralling into chaos, this is the episode to watch.

David Draper is the Data Science Manager at IRIS Software Group, where he leads a high-performing team at the intersection of engineering, analytics, and emerging tech. With a background in education and a passion for quantum computing, David brings a unique lens to the challenges of modern data strategy.

Timestamps and Takeaways
✔️04:42 – From Classroom to Data Strategy
David’s journey from education to enterprise data science
→ Leverage curiosity and teaching mindset to lead technical teams with clarity and empathy

✔️09:15 – Why Quantum Thinking Isn’t Just for Physicists
How quantum logic helps IRIS reimagine compute and decision-making
→ Adopt non-linear thinking to structure smarter, more adaptable infrastructure

✔️14:08 – Designing Modular Systems for Innovation Without Risk
How IRIS builds infrastructure that allows safe experimentation
→ Create sandbox-style systems to test and deploy without affecting BAU

✔️19:22 – Ring-Fencing Innovation Time Inside a Busy Enterprise
Balancing research and delivery with “go wide, then narrow” phases
→ Allocate structured exploration time to prevent constant firefighting

✔️24:50 – The Real ROI of Explainable AI
Why clarity builds trust and momentum across the business
→ Choose tools your stakeholders understand to drive adoption and reduce resistance

✔️30:30 – Building Teams That Experiment Responsibly
How culture, structure and trust shape IRIS’s approach to innovation
→ Foster autonomy while staying aligned to business goals

Why Listen

This episode is for data leaders, architects, and CTOs who want to scale responsibly, embed innovation, and prepare for what’s next—without being distracted by hype. It’s a tactical, grounded conversation that offers immediately applicable ideas.

🔔 Subscribe to Data Matas for more real conversations with data leaders.
💬 What’s one lesson from David’s approach you’d apply to your own team? Drop your thoughts in the comments.
🎧 Available on YouTube, Spotify, Apple Podcasts, and all major platforms.

Links
🔗 David Draper on LinkedIn: https://www.linkedin.com/in/david-draper-b715aa46/
🔗 IRIS Software Group: https://www.linkedin.com/company/iris-software-group/
🌐 Listen to more episodes: https://pod.link/1763791020
📚 Matatika resources: https://www.matatika.com/library/
📺 Watch more episodes: https://www.matatika.com/podcasts/

What Crypto Data Teams Do Differently with Emily Loh

Episode Snapshot
In this illuminating conversation, Emily Loh reveals how leading a data team in the volatile crypto space has forced innovations in resource allocation, strategic prioritisation, and AI implementation that data leaders across all industries can apply for greater impact.

Guest Introduction
Emily Loh serves as Director of Data at MoonPay, where she leads a 15-person team spanning data engineering, data science, and machine learning in a fast-evolving crypto environment. With previous experience at Coinbase and a surprising background in literature rather than computer science, Emily brings a unique storytelling perspective to data leadership that emphasises business outcomes over technical outputs.

Conversation Highlights
The discussion takes an unexpected turn when Emily reveals her humanities background, noting that her literature studies prepared her surprisingly well for data leadership: "I never thought I'd be in data, but specifically I studied literature. And in my career right now, I'm just like, 'oh, this is just really just storytelling.'" This insight evolves into a thoughtful exchange about how effective data work requires understanding human needs first, with technical implementation second.

A particularly candid moment occurs when Emily confesses, "Full disclosure, I am a formed people pleaser," sparking an authentic discussion about the struggle many data leaders face in saying "no" to low-value requests. Aaron and Emily explore how this seemingly simple act requires both courage and strategic frameworks to implement effectively.

Actionable Insights

1. Implement the 20/40/40 resource allocation framework - Emily's team divides their time into 20% BAU (business as usual), 40% building, and 40% research, creating space for innovation even during challenging market periods. Practical Application: Start by auditing your team's current time allocation, then gradually shift toward this balanced model using clear opportunity sizing frameworks to evaluate potential projects.

2.Break free from the "service trap" by transforming request handling - Instead of immediately building requested dashboards, train your team to ask "What decisions are you trying to make?" to focus on outcomes rather than outputs. Practical Application: Develop a structured intake process that guides stakeholders toward better request formulation and establishes clear value criteria for accepting work.

3. Use AI strategically to eliminate team drudgery - At MoonPay, tools like Cursor help automate tedious tasks such as YAML file management, freeing analyst time for strategic work. Practical Application: Survey team members about their most tedious regular activities, then select appropriate AI tools that can automate these specific tasks while measuring success through time savings.

4. Design data systems for uncertain futures - In crypto's rapidly changing landscape, Emily's team builds flexible architectures that can adapt to regulatory shifts and market changes. Practical Application: Implement modular data models that can evolve without complete rebuilds while maintaining strong data quality foundations that support agility regardless of specific technologies.

Industry Context
This conversation arrives at a critical inflection point for data leaders across industries, as generative AI promises transformation while many teams still struggle with the fundamental challenge of delivering strategic value rather than reactive reporting. Emily's experience navigating crypto's extreme volatility provides a stress-tested framework applicable to any data team facing resource constraints and rapid change.

Why Listen
This episode is essential for mid to senior-level data leaders who feel trapped in reactive work cycles and are seeking practical frameworks to increase their strategic impact. Whether you're in a traditional enterprise struggling with legacy approaches or a high-growth startup trying to balance immediate demands with future needs, Emily's battle-tested insights provide immediately applicable strategies for transformation.

Episode Details

Length: 37 minutes
Release Date: March 2025
Episode Number: #127

This conversation with Emily Loh offers rare insight into how crypto's extreme conditions have forced innovation in data team management—innovations that can give any data leader a competitive advantage in today's rapidly evolving landscape.

Are you unknowingly overspending on cloud data infrastructure?

Many businesses migrate to the cloud expecting cost savings and efficiency, but hidden costs, vendor lock-in, and inefficient ETL processes often result in ballooning expenses. Without a strategic approach, organisations risk wasting budget on unnecessary compute, storage, and manual data management.

What You’ll Learn in This Episode

In this episode of Data Matas, host Aaron Phethean speaks with Jon Hammant, Head of Compute for UK & Ireland at AWS, about the true cost of scaling data infrastructure and how businesses can optimise cloud spend. Jon shares his insights on avoiding pricing traps, reducing data migration costs, and leveraging AI-driven automation to improve efficiency.

Key Insights & Timestamps

1.  - Reducing unnecessary data syncing to cut costs
Real-time syncing is often overused, leading to excessive compute costs. Discover how batch processing can reduce ETL expenses by up to 50%.

2.  - Conducting cloud audits to eliminate waste
Many organisations pay for idle compute and unused storage without realising it. Learn how to audit cloud usage and remove unnecessary expenses.

3. - Avoiding vendor lock-in and costly renewal contracts
Row-based ETL pricing can trap businesses into increasing costs as data volumes grow. Find out how switching to usage-based pricing can provide more control over cloud spend.

4.  - Automating data management with AI
Manual ETL processes drain resources and increase operational costs. Learn how AI-driven automation can streamline workflows, reduce errors, and improve efficiency.

About Jon Hammant

Jon Hammant is the Head of Compute for UK & Ireland at AWS, where he helps businesses optimise cloud infrastructure, modernise data strategies, and reduce operational costs. With extensive experience in AI, cloud computing, and high-performance networking, Jon has worked with some of the world’s largest enterprises to scale data infrastructure efficiently without unnecessary spend.

Subscribe & Join the Conversation

If you're looking to optimise cloud spend, reduce migration costs, or navigate vendor pricing models, this episode is for you.

🔔 Subscribe to Data Matas to get the latest insights on cloud cost optimisation, data infrastructure, and AI-driven efficiency.

Resources & Links
📌 Jon Hammant on LinkedIn: https://www.linkedin.com/in/jhammant/
📌 Matatika Podcast & Resources: https://www.matatika.com/podcasts/s2e1-scaling-your-data-infrastructure-with-jon-hammant

In this Season 1 Highlights episode, we break down seven proven strategies that help teams cut costs, improve data quality, and scale smarter.

We’ve spoken with data leaders, engineers, and BI experts who have tackled real-world challenges—from legacy systems and broken workflows to AI risks and siloed teams. This episode is your actionable playbook to making your data work better.

Key Takeaways:

✅ Fix data chaos and create a single source of truth – Jessica Franks (Not On The High Street) shares how she aligned business and tech using Wardley Maps.
✅ Rebuild trust in business intelligence – Joe Wright (CitySprint) explains how they solved reporting inconsistencies by consolidating systems.
✅ Scale smarter without overcomplicating – Stéphane Burwash (Potloc) shows how open-source tools and a data champions programme transformed their approach.
✅ Why ‘good enough’ beats perfection – Bethany Lyons explains how streamlining data pipelines saves time without sacrificing quality.
✅ Make data quality everyone’s job – Adam Dathi (MVF) reveals how cross-team collaboration fixes unreliable reporting.
✅ Using real-time data for better decision-making – Nick Bromley shares how transport data integration is driving smarter city planning.
✅ AI without the risk – Murtaza Kanchwala (Amplify Capital) details how his team successfully implemented AI while staying compliant.

🚀 Whether you're a Head of Data, CTO, BI Manager, or Data Engineer, these practical insights will help you fix inefficiencies, scale with confidence, and build a high-impact data team.

🎧 Listen now and take your data strategy to the next level and 📩 Subscribe for more insights

In this conversation, Murtaza Kanchwala from Amplifi Capital discusses the integration of Generative AI (Gen.AI) in financial services, sharing insights on early applications, challenges faced, and the importance of regulatory compliance. He emphasizes the need for effective team organization and the selection of appropriate AI models to enhance productivity and efficiency. The discussion also touches on the future of AI in finance, predicting the emergence of personalized AI assistants and the ongoing evolution of AI technologies.

Takeaways

Murtaza has been involved in financial services since 2014.

Gen.AI applications started with content generation in 2021.

Feedback from users was gathered through personal discussions.

AI hallucinations posed significant challenges in early implementations.

Internal AI solutions were prioritized before customer-facing applications.

Testing AI requires a different approach than traditional software.

The maturity of Gen.AI use cases is improving over time.

Amplifi Capital is building an AI Matrix platform for Gen.AI use cases.

Choosing the right LLM is crucial for specific use cases.

Regulatory compliance is essential in financial services AI applications.

Sound Bites

"AI is delivering something artificially."

"The landscape is moving faster than we think."

Chapters

00:00 Introduction to Gen.AI in Financial Services

06:27 Early Applications of Gen.AI in Finance

11:07 Maturity of Gen.AI Use Cases

16:51 Building the AI Matrix Platform

22:48 Regulatory Landscape for AI in Finance

29:37 Building Effective Squads in AI Projects

35:18 Future of AI in Financial Services

In this conversation, Aaron Phethean and Nick Bromley discuss the evolution and importance of transport data, particularly focusing on the integration of real-time data and mobile phone data into transport planning. They explore the challenges of data collection, the role of AI and big data in optimizing transport systems, and the future of transport data with an emphasis on privacy and security concerns.

Takeaways

• Good transport planning requires both short-term and long-term data analysis.
• Buses provide the most flexible capacity in public transport systems.
• Real-time data is crucial for understanding current transport demands.
• Historical data often fails to reflect current population movements.
• Mobile phone data can significantly enhance transport planning accuracy.
• Data collection methods must evolve to include modern technology.
• AI and big data can process vast amounts of transport data effectively.
• Privacy concerns must be addressed when using personal data for transport planning.
• Transparency in data usage is essential for public trust.
• The future of transport data relies on secure and anonymized data sharing.

Sound Bites

"What does good look like for transport data?"

"It's ludicrous to use 1920s data."

Chapters

00:00 Introduction to Innovation in Transport Data

03:28 The Importance of Buses in Urban Mobility

04:06 Understanding Transport Data Needs

06:54 The Role of Mobile Phone Data in Transport Planning

09:35 Challenges and Innovations in Data Collection

12:29 The Future of Data Privacy and Public Good

15:21 AI and Big Data in Transport Decision Making

18:15 Conclusions and Future Directions for Transport Data

This episode with Adam Dathi is a must-listen for anyone looking to turn data into a strategic powerhouse within their business.  Adam shares practical insights into how data teams can work seamlessly with other departments for maximum business-wide impact and gives his take on the future of AI-driven data analysis. Aaron and Adam discuss the critical role of reliable sources and governance, but also observe that this is not an isolated issue for a specific team, but a company-wide responsibility if one is looking to harness the true power of data for their business.

Takeaways

• Data quality is a company problem, not just a data problem.
• The sophistication of data teams depends on company size.
• Data governance is a hidden long-term investment.
• AI can't generate ideas without human direction.
• Data maturity impacts decision-making effectiveness.
• Data teams need to interface better with the rest of the company.
• The value of data is tied to decision-making outcomes.
• Standardizing terms can improve data governance.
• Data quality issues often stem from company culture.
• Investing in data governance can yield hidden benefits.

Titles

• AI and the Future of Data Analysis
• Unlocking the Power of Data in Business
• Sound Bites
• "Data quality is a company problem, not just a data problem."
• "The sophistication of data teams depends on company size."
• "Data governance is a hidden long-term investment."

Chapters

• 00:00 Introduction to Data in Marketing
• 04:43 The Role of Data in Strategic Decision-Making
• 07:34 MVF: An Integrated Media and Marketing Company
• 10:45 Challenges in Data Technology and Team Dynamics
• 13:52 Data Quality vs. Company Culture
• 16:49 The Importance of Data Governance
• 19:55 AI's Role in Data Analysis and Decision-Making
• 22:39 The Future of Data and AI in Business
• 25:51 Conclusion and Reflections on Data's Impact
• 49:59 Wrapup
• 49:59 Final Thoughts and Reflections