The Data Mix Podcast · 2025-01-20 · 59 min
Key moments - from our scoring
Substance score
63 / 100
Five dimensions, 20 points each
Ben Rogojan, known as the Seattle Data Guy, traces his evolution through healthcare analytics, SQL Server-based data warehousing, and several years at Facebook before transitioning to independent consulting and content creation. The episode digs into what data engineering actually means across different organizational contexts - contrasting the infrastructure-heavy work at most companies against the abstracted, platform-focused role at Facebook. At typical enterprises, data engineers must handle foundational infrastructure challenges: spinning up Airflow orchestration from scratch, integrating seven incompatible ERP systems, managing sprawling data silos, and building data models while fighting constant technical friction. Facebook removed most of that friction, allowing engineers to focus on impact rather than infrastructure. Ben emphasizes the critical distinction between platform-level data engineering (building foundations), infrastructure data engineering (creating core data models that change slowly), and analytics-layer work (building business value). He also touches on emerging challenges like breadth requirements across Python, orchestration tools, and cloud platforms, plus the growing politicization of tech companies and regulatory pressures that data teams must navigate.
Ben found it rewarding to work alongside exceptionally smart people and appreciated the removal of technical friction through company-built infrastructure, freeing teams to focus on impactful work rather than solving foundational problems like data integration or Airflow management.
At Facebook, data infrastructure is pre-built with integrated datasets and certified data cataloging, whereas typical companies must build these foundations themselves - managing multiple incompatible ERP systems, spinning up Airflow orchestration, and handling data silos.
Platform infrastructure (managed by specialists), core data models that represent business entities and change slowly (treated like software), and the analytics layer where business value is built through visualizations and use cases.
Data engineers increasingly need breadth across Python scripting, cloud platforms, orchestration tools like Airflow, data modeling, and deployment/release processes - not just depth in a single tool.
Non-FANG companies lack mature platforms and must integrate legacy systems, manage data silos across incompatible ERPs, and make trade-offs between expensive architectural fixes and quick solutions that limit long-term value.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode contains solid practitioner insights about data engineering across different company scales (Facebook vs startups) and specific challenges like data integration, governance, and orchestration. However, much of the discussion is somewhat general (tools overview, career advice) and the insights often lack the depth and specificity that would elevate this to 15+. The bonus segment on acquisitions adds some value but feels rushed and surface-level.
the data is integrated you know because the way they've developed their product it's it's also kind of ground up it's digital um there's a lot of benefits that you have there versus a lot of companies have like seven ERP systems none of them talk to each other
if you don't respect that as code like if you don't treat it like as a system that you're now developing that in the future can cause unknown issues or have unknown issues um there's a huge risk
Ben articulates some useful distinctions (foundational vs applied data engineering, three-layer architecture) but these are not particularly novel frameworks. The consultant's perspective on client selection and business reality is practical but fairly standard. The bonus acquisition commentary adds little original analysis. The episode largely confirms existing thinking rather than challenging conventional wisdom about data engineering.
you have to build the what I consider kind of infrastructure still but it's you know that data layer that is your core data model which represents uh your company's business
I'm writing an article uh a little bit about that about like thinking about data engineering Beyond Silicon Valley because I think we read a lot of Articles uh from Silicon Valley and and and you know see the cool technology that they have
Ben is a genuine practitioner with meaningful experience: ~10 years in data, concrete tenure at Facebook, healthcare analytics background, and active consulting work. He's not a pure thought-leader or career podcaster. However, he's not at the CEO/CTO level of a major company, and his primary visibility comes from content creation rather than being known for a specific high-stakes achievement, which limits the score.
I worked at a hospital for a little bit doing kind of SQL Server ssas uh data warehousing and then work for a healthcare analytics company
I did Facebook for a few years uh again kind of working in data engineering and then finally I've been Consulting kind of uh the whole time
The episode is frustratingly light on concrete examples. Ben mentions Facebook, healthcare companies, and some tools (Airflow, Tableau, Snowflake) but rarely with specific metrics, case studies, or measurable outcomes. The consulting work is described generically (data migration, infrastructure setup) without named clients or detailed project results. The bonus segment names Upsolver/Databricks/DBT but lacks numbers or concrete impact data.
I remember having to go through some people's pipelines um like some machine learning Engineers pipelines and it was like thousands and thousands of lines um and and and that was a little bit hairy
I I've had a few companies that I've worked with and partnered with where they usually bring me in on the on the flip side right they're like hey we're doing the Salesforce migration
The hosts ask competent, relevant questions and Ben provides thoughtful answers, but the conversation rarely pushes deeper with sharp follow-ups or productive disagreement. Questions are mostly open-ended invitations to elaborate rather than probing specific claims. The dynamic is friendly but lacks the intellectual friction that would signal genuine interrogation. The bonus segment feels tacked-on and rushed.
what can you say about that in terms of the definition of like there's kind of how you handle data engineering and there's kind of foundational data engineering so what are some of the differences there
you mentioned HubSpot sort of HubSpot to Salesforce is um and the and the data behind it is that a is thatan area that you would touch
Computed from the transcript - who did the talking, and the words that came up most.
Send a text And just like that, we're back with a candid discussion on the state of data engineering in 2025! Join us as we sit down with Ben Rogojan, aka the Seattle Data Guy, for a sharp analysis of the consulting landscape. With over a decade of experience in data infrastructure and analytics, Ben offers a thoughtful look at: • The evolution of data engineering roles and skills • Challenges faced by consultants in today's market • Navigating the transition from full-time to consulting work • Essential tools and technologies shaping the field Don't miss this insider's perspective on cloud migration, AI integration, and the future of business intelligence. Whether you're a seasoned pro or considering a career shift, Ben's insights will help you stay ahead of the curve in data engineering. Tune in for expert advice on building your personal brand, attracting clients, and thriving in the ever-changing world of data. This episode is packed with actionable tips and industry trends you won't want to miss!
Transcribed and scored by The B2B Podcast Index.
[Music] your fix of the best guests from the world of data and analytics this is the data mix with Brian buen and George pon [Music] Brian George we're back high five oh oh gosh I'm might practiceing this way though it's always the opposite right yeah yeah nice one man good to see you how are you ah pretty good thank you I was up early this morning um it was a half past 4 start to go from my cold wet home in Scotland down to London the Big Smoke as we call it yeah I I think we should have like a like a we's Wally guess weGeorgia type of segment as well because like you're never in the same place twice mate but it's always always a pleasure to have you here and no 4 a.
m.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
How B2B Brands Use AI for Churn PredictionThe Growth Operator with Fexingo · on Data warehousing83 / 100
The new social with Rachel KartenThe Rebooting Show · on Facebook82 / 100
Who really owns the Bill of Materials?AI Across The Product Lifecycle Podcast · on ERP systems80 / 100
Redefining Hospitality Through Tech - Tanya Pratt - Defining HospitalityDefining Hospitality · on Data silos78 / 100
Why Your Planning System Can't Help You ChangeSpeaking of Supply Chain · on ERP systems76 / 100
[REPLAY] Tinder, TripAdvisor, and more: Universal Product LessonsProduct Rebels · on Facebook75 / 100TDM S3 Ep10 - Qlik Connect 2025 with Mike Capone - Navigating the Agentic Revolution58 / 100TDM S3 Ep9 - Beyond the Buzzwords: Rethinking Our Approach to Data and AI Literacy76 / 100TDM S3 Ep8 - Women Who Qlik: Driving Data Diversity62 / 100TDM S3 Ep7 - Taylor Desseyn: Building Trust in Tech Communities80 / 100TDM S3 Ep6 - Deepak Prasad: Catching the Data Wave Before It Breaks72 / 100