DataTalks.Club Podcast Wiki

Guide

Data Engineering Podcast: DataTalks.Club Episodes to Start With

A podcast-backed listening guide to DataTalks.Club episodes on data engineering fundamentals, tools, platforms, DataOps, careers, freelance work, streaming, Data Mesh, and governance.

Related Wiki Pages

Data Engineering Data Engineering Platforms Data Engineering Roadmap Modern Data Stack DataOps Data Quality and Observability Batch vs Streaming Data Mesh Data Governance Freelance

If you’re looking for a data engineering podcast, use this guided path through DataTalks.Club episodes. It starts with pipelines, warehouses, orchestration, and DataOps. It also covers platforms, streaming, governance, and careers. Freelance work has its own section. Start with role definitions, then move into tooling and operating problems, then into career and consulting episodes.

Use Data Engineering for the reference layer behind the listening path. Pair it with Data Engineering Platforms, Data Engineering Roadmap, and Modern Data Stack.

Listening Map

Use these sections to choose the next cluster to listen to.

Fundamentals

Start here if you want the base vocabulary because these episodes explain what data engineers do. They show where engineers hand work to analysts and data scientists, and why a pipeline needs ownership beyond a scheduled script.

Start with Data Team Roles Explained, the archive’s first role map. Around 13:58, data engineers make necessary data available in usable form. Around 30:01, the episode separates data preparation from machine learning work.

Use Big Data Engineer vs Data Scientist for older big-data vocabulary. Roksolana Diachuk starts from ETL pipelines, HDFS/S3, Impala, and Spark performance work. She then compares data engineering with data science, ML deployment, and analyst-facing interfaces.

Data Engineering Tools and Modern Data Stack is the foundation episode for modern-stack vocabulary. Natalie Kwong walks through ETL, ELT, ingestion, and transformations. She also covers data marts, warehouses, and lakes. Later sections cover orchestration, reverse data flows, CDC, and schema evolution.

For written companions, use Data Engineering and Fundamentals of Data Engineering.

Tools and Modern Stack

The tool episodes are useful because guests keep tying products back to data flow. They discuss ingestion, storage, transformation, and orchestration. Activation and cost come up as operating concerns too.

Data Engineering Tools and Modern Data Stack is the cleanest starting point:

Modern Data Engineering updates that map for lakehouse and AI-era work. Adrian Brudaru covers Apache Iceberg, Delta Lake, and Hudi. He also discusses DuckDB, metadata catalogs, orchestration, and streaming. He also discusses AI-assisted pipeline work. Around 44:42, he warns against vendor-led tool choices.

Around 41:06, he brings the learning path back to SQL and Python, plus requirements and portfolio evidence.

In Data-Led Growth, Event Tracking, and Reverse ETL, Arpit Choudhury shows the product and growth side of the stack. He connects tracking plans to warehouse work, then moves into transformations and reverse ETL. Customer data platforms and operational analytics come next. Use this episode when you want to understand why a warehouse table isn’t the end of the data flow.

For deeper tool comparisons, use these pages:

DataOps and Quality

Data engineering work breaks in ordinary ways because files arrive late, schemas change, and joins duplicate records. Dashboards go stale, and downstream models can train on broken inputs. The reliability episodes treat these failures as operating problems, not as one-off bugs.

DataOps for Data Engineering connects pipeline work to automation, observability, CI/CD, and regression tests. It also covers realistic test data, deployment confidence, and recovery. Christopher Bergh makes the case that data teams need operating practices, not only orchestration.

In Data Observability Explained, Barr Moses anchors the quality layer. Around 16:38, she defines the five pillars as freshness and volume, distribution and schema, and lineage. Later sections cover schema-change incidents and SLAs. They also cover runbooks, alerting, and ownership. Use the episode for the difference between detecting a failure and finding the cause.

Data Engineering for Fraud Prevention shows data quality under time pressure. Angela Ramirez discusses feature pipelines, daily batches, real-time scoring, and Great Expectations. She also covers schema changes, logs, runbooks, and debugging. It’s the most concrete episode here for seeing how bad data can affect a live operational decision.

For the written synthesis, use these pages:

Platforms and Team Scale

Data engineering becomes platform work when many teams need reliable data without waiting for one central engineer to handle every request. The archive keeps a useful tension here. Self-service helps teams move faster, but only when shared conventions, ownership, and monitoring exist.

Scaling Data Engineering Teams is the core platform episode. Mehdi OUAZZA describes scale-up data work, self-service platforms, and onboarding. He also covers Airflow conventions, playbooks, and Kafka. Schema registries and data contracts are part of the same platform discussion.

Around 17:22, he explains the platform anatomy. Around 23:26, he connects Kafka to schemas and contracts rather than treating it as a standalone streaming tool.

Data Engineering Leadership and Modern Data Platforms adds the management view. Rahul Jain uses his ETL-to-platform-management path to discuss stakeholder management, team prioritization, and data reconciliation. He also discusses GDPR controls, lineage, and the shift from ETL to ELT.

FinOps for Data Engineers adds the cost dimension. Eddy Zulkifly ties cloud cost work to BigQuery, storage tiers, reservations, and tagging. He also covers accountability, forecasting, and managed-versus-custom tradeoffs.

For the reference layer, use these pages:

Careers and Learning Paths

The archive’s career advice is consistent: SQL and Python matter before broad tool collecting. Data modeling, tests, and requirements gathering matter too. You need finished projects more than shallow exposure to Spark, Kafka, Kubernetes, or a lakehouse stack.

Build a Data Engineering Career is the main beginner-curriculum episode. Around 23:35, Jeff Katz names Python, SQL, and cloud fundamentals as core skills. Around 38:05, he explains why a junior-focused path may skip Spark, Kafka, and Kubernetes until the basics are solid.

Data Engineer Career in 2026 is useful for role definition. Slawomir Tulski separates platform data engineering from product-facing data engineering, questions overbuilt stacks, and treats cost-aware judgment as a senior signal. Use the portfolio sections around 57:35 and 1:04:42 when deciding what kind of project proves readiness.

Data Engineering Job Prep and Interview Guide and Data Engineering Job Search Story are useful when you want hiring and interview context. Jeff Katz covers Python, SQL, Docker, and Airflow. He also covers portfolio work, interviews, and application strategy.

Gloria Quiceno gives a job-search case study from neuroscience research into analytics and data engineering. The episode covers Docker, Airflow, and AWS. It also covers application tracking and a custom portfolio project.

For written guides, use these pages:

Freelance and Consulting

The archive has a strong independent-work thread because several data engineering episodes involve freelancers, consultants, and founders who turned repeated client problems into reusable tools.

Freelance Data Engineering Playbook is the main episode. Adrian Brudaru covers pricing, occupancy, agencies versus direct clients, and repeat business. He also covers scoping, discovery spikes, written scope documents, and volatile schemas. Remote work, reusable assets, and client expectations come up too.

Use 11:36 for early projects that involved legacy cleanup and Airflow. Use 31:43 for scoping through spikes and written scope documents. Use 46:17 for reusable portfolio assets.

Modern Data Engineering also helps here because Adrian connects freelance experience to tool building and reusable ingestion patterns. Listen to the early career section around 3:10 and the DLT discussion around 4:03 if you want the bridge from consulting pain to product work.

Data Engineer Career in 2026 adds a senior consulting perspective. It helps clarify when clients need a platform specialist, product data engineer, or consultant who can diagnose the operating constraint before choosing a tool.

For written follow-up, use these pages:

Batch and Streaming

The podcast archive doesn’t treat streaming as a maturity badge. Guests use streaming when the decision needs low latency. They use batch, scheduled jobs, or micro-batches when the business can wait and the simpler system is easier to operate.

Modern Data Engineering gives the most recent vocabulary. Around 51:19, Adrian Brudaru compares micro-batching and Kafka, then discusses SQS and Flink. Name the SLA before naming the tool.

Scaling Data Engineering Teams explains the organizational side of streaming. Mehdi OUAZZA treats Kafka as part of a platform with schemas, registries, contracts, and guidelines. Consumer ownership matters too.

Data Engineering for Fraud Prevention is the clearest applied example. Angela Ramirez combines daily feature batches with real-time scoring because fraud systems need to support decisions during a transaction.

From Notebooks to Production and DataOps 101 for Scaling Data Platforms also compare batch and streaming. Andreas Kretz covers queues, event pipelines, and model-serving tradeoffs. Lars Albertsson covers micro-batching, streaming, dependency management, and reproducibility from a DataOps platform focus.

For written comparison, see Batch vs Streaming and Streaming.

Data Mesh and Governance

Data engineering work becomes governance work when teams need shared meaning, access control, and ownership. Quality signals, retention, and policy automation belong in that work too.

Data Mesh Implementation is the core Data Mesh episode. Zhamak Dehghani explains domain ownership and data products, then moves into contracts, quality guarantees, and SLAs. Self-serve platforms and federated governance come next.

Use 13:20 for contracts and domain ownership, 31:05 for metadata and interoperability, and 41:58 for self-serve platforms and federated governance.

How to Build Data Governance in the Cloud is the practical governance episode. Jessi Ashdown and Uri Gilad define governance beyond security and PII. They cover classification, catalogs, policies, and data stewards. They also cover data quality signals, retention, and freshness.

They discuss purpose-based access, automation, access workflows, and ROI later.

DataOps 101 for Scaling Data Platforms adds a useful caution: decentralization creates ownership and governance risk if teams don’t have mature operating practices. Pair it with the Data Mesh episode if you want the tradeoff between domain autonomy and shared standards.

For written synthesis, use these pages:

Build a Listening Path

Choose the path based on the question you’re trying to answer.

The best way to listen is to pair one concept episode with one applied episode. For example, listen to Data Engineering Tools and Modern Data Stack, then listen to Data-Led Growth. That pairing shows how warehouse data moves back into customer-facing work. Or listen to the batch-versus-streaming material, then listen to the fraud-prevention episode to see why latency changes the architecture.

Use these pages after you choose an episode cluster.