A path per stack, not a bookmark dump

YouTube

• Snowflake Developers

  Official product walkthroughs, Cortex, and release recaps.

• Snowflake Inc.

  Summit keynotes and platform strategy — useful for Cortex and Iceberg context.

• freeCodeCamp — Snowflake for beginners

  Long-form free intros when you want one sitting, not a playlist of clips.

• Snowflake Cortex / Gen AI sessions

  Analyst, Search, and COMPLETE — watch 2025+ only; the UI moved.

Courses

• Udemy — Snowflake topic

  Filter for recent, high-review courses. Prefer ones that cover ACCOUNT_USAGE and warehouses, not only UI clicks.

• Snowflake University (official)

  Hands-on labs that match SnowPro Core and Specialty: Gen AI.

• Coursera — Snowflake

  Partner courses if you want a certificate path next to SnowPro.

• DataLane SnowPro Gen AI practice

  Original MCQ in the shape of the specialty exam — not a dump.

Books

• Snowflake Cookbook (Packt)

  Task-shaped recipes: load, share, govern. Use as a lookup, not a novel.

• Snowflake: The Definitive Guide (O'Reilly)

  Architecture, cost, and governance in one narrative. Pair with current docs for Cortex.

• Fundamentals of Data Engineering (Reis & Housley)

  The lifecycle book. Read the warehouse and orchestration chapters before another Snowflake tutorial.

YouTube

• dbt Labs

  Coalesce talks, Coalesce keynotes, and how the product actually changed.

• dbt Learn

  The official fundamentals path, still the right first 4 hours.

Courses

• dbt Learn (official, free)

  Fundamentals then advanced. Do this before any paid Udemy clone.

• Udemy — dbt

  Useful after official Learn if you want extra incremental / Slim CI labs.

Books

• Analytics Engineering with dbt (forthcoming / O'Reilly tracks)

  Pair with the official docs on incremental, snapshots, and contracts.

• The Data Warehouse Toolkit (Kimball)

  Grain and facts. dbt does not replace dimensional modeling.

YouTube

• Apache Airflow

  Summit talks and release walkthroughs — watch Airflow 3 sessions first.

• Astronomer

  Clear DAG design talks; treat product pitches as optional.

Courses

• Udemy — Airflow

  Pick a course updated after Airflow 2.7+. Skip anything that still teaches BashOperator-only DAGs.

• The Astronomer Academy

  Free modules on deferrable operators, datasets, and local dev.

Books

• Data Pipelines with Apache Airflow (O'Reilly)

  Still the best narrative book. Pair with current docs for Airflow 3 APIs.

YouTube

• Databricks

  DAIS keynotes, Delta, Unity Catalog, and Mosaic/AI sessions.

• Advancing Analytics

  Practitioner Spark/Databricks walkthroughs with real clusters.

Courses

• Databricks Academy (official)

  The path that matches DEA / professional exams.

• Udemy — Databricks

  Use for extra Spark labs; ignore courses that skip Unity Catalog.

Books

• Learning Spark, 2nd ed. (O'Reilly)

  The Spark mental model. Databricks is an implementation of this.

• Delta Lake: The Definitive Guide (O'Reilly)

  Transaction log, OPTIMIZE, and time travel without slideware.

YouTube

• Apache Spark

  Summit talks on AQE, shuffle, and Structured Streaming.

• Databricks Spark tutorials

  PySpark DataFrame API in a real notebook environment.

Courses

• Udemy — PySpark

  Filter for courses that teach the Spark UI, not only df.show().

Books

• High Performance Spark (O'Reilly)

  Shuffle, join strategy, and memory — the interview book.

• Learning Spark, 2nd ed.

  Start here if you have not run a job that failed on shuffle.

YouTube

• Apache Kafka

  Kafka Summit talks on transactions, KRaft, and Connect.

• Confluent

  Schema Registry, ksqlDB, and Flink — know what is Kafka vs Confluent.

Courses

• Udemy — Kafka

  Prefer courses that run a local cluster and a consumer group, not slides only.

• Confluent Developer

  Free courses on fundamentals and stream processing.

Books

• Kafka: The Definitive Guide, 2nd ed. (O'Reilly)

  Partitions, replication, consumers. Required reading before a streaming design interview.

• Designing Data-Intensive Applications (Kleppmann)

  Logs, replication, and why Kafka looks the way it does.

YouTube

• freeCodeCamp — SQL

  A complete beginner-to-joins pass if you are rusty.

• Mode Analytics SQL tutorial

  Written + interactive. Better than most video for windows.

Courses

• Udemy — SQL

  Pick window-function heavy courses. Skip 40-hour SELECT * tours.

• DataLane SQL Playground

  Practice the same windows and joins on sample orders in this tab.

Books

• SQL for Data Analysis (O'Reilly)

  Windows, cohorts, and the queries analysts actually write.

• Joe Celko's SQL for Smarties

  Set thinking. Overkill for day one; gold for interview edge cases.

• The Art of SQL (Faroult)

  How to think in sets when an interview whiteboard is empty.

YouTube

• mCoding / real Python talks

  Prefer 2024+ videos that mention Arrow and polars, not only pandas 1.x.

• ArjanCodes

  Testing and structure for scripts that become pipelines.

Courses

• Udemy — Python for data

  A short pandas/polars course beats a 60-hour “everything Python” mega-class.

Books

• Python for Data Analysis, 3rd ed. (McKinney)

  pandas from the author. Still the reference; know where polars diverges.

• Fluent Python, 2nd ed. (Ramalho)

  Iterators, typing, and concurrency — the parts pipeline code actually uses.

YouTube

• AWS Developers

  Glue, Redshift, Athena, and Kinesis sessions from re:Invent.

• AWS Events

  Keynotes when a service actually GA'd, not just a blog title.

Courses

• AWS Skill Builder (official)

  Free digital training; paid labs if you want the DEA path.

• Udemy — AWS Data Analytics / DEA

  Exam-shaped practice. Cross-check every service against current AWS docs.

Books

• Data Engineering with AWS (Manning / Packt titles)

  Architecture over click-ops. Pair with the Well-Architected Data Analytics lens.

• Designing Data-Intensive Applications (Kleppmann)

  Why S3 + a log looks like Kafka. Read this before another Glue tutorial.

YouTube

• Microsoft Azure

  Fabric, Synapse, and ADF sessions — watch Fabric ones from the last 12 months only.

• Guy in a Cube

  Power BI + Fabric serving layer, which is where most Fabric cost shows up.

Courses

• Microsoft Learn — Fabric / DP-203

  Official modules. Start here before Udemy.

• Udemy — Microsoft Fabric / Azure data

  Useful labs; ignore anything that still treats ADF as the whole platform.

Books

• The Definitive Guide to Azure Data Engineering (search current ed.)

  Synapse vs Fabric changes fast — prefer 2025+ editions.

• Fundamentals of Data Engineering (Reis & Housley)

  The lifecycle still applies when the logo says Fabric.

YouTube

• Google Cloud Tech

  BigQuery, Dataflow, and Pub/Sub deep dives from Cloud Next.

Courses

• Google Cloud Skills Boost

  Official quests for PDE. The BigQuery ones are worth the time.

• Udemy — BigQuery / GCP data

  Supplement labs; billing models change — verify in current docs.

Books

• Google BigQuery: The Definitive Guide (O'Reilly)

  Slots, partitions, and the cost model. Still the book to own.

• Streaming Systems (Akidau et al.)

  Watermarks and event time — the Dataflow mental model.

YouTube

• DuckDB

  Official talks on the vectorized engine and extensions.

Courses

• DuckDB docs + tutorials

  The official guides beat any paid course today.

Books

• DuckDB in Action (Manning)

  Local analytics and when not to stand up a warehouse.

• Python for Data Analysis, 3rd ed.

  When the local file is a DataFrame first and DuckDB second.

YouTube

• Made With ML

  Feature stores and training jobs as pipelines — the DE half.

• Databricks Mosaic / ML sessions

  Unity Catalog + feature tables, not a notebook-only demo.

Courses

• Udemy — MLOps

  Pick courses that version data and models. Skip ones that only wrap Flask.

• DeepLearning.AI MLOps

  Coursera specialization if you want the long path.

Books

• Designing Machine Learning Systems (Huyen)

  Data-centric ML. Required if you own the feature table.

• Reliable Machine Learning (O'Reilly)

  Production failures — closer to DE on-call than a Kaggle book.

YouTube

• DeepLearning.AI

  Short RAG and agent courses that stay practical.

• Snowflake Cortex sessions

  Warehouse-native AI — closer to this site than generic chatbot demos.

Courses

• DeepLearning.AI short courses

  RAG, evals, and agents in hours, not months.

• Udemy — LangChain / RAG

  Pick courses that talk about evals and cost per row, not only hello-world chains.

Books

• Designing Machine Learning Systems (Huyen)

  Data-centric ML. The DE half of the MLOps conversation.

• AI Engineering (O'Reilly, 2025+)

  Application patterns: retrieval, eval, and ops — not model training.

YouTube

• Kimball Group talks / Data Modeling Zone

  Grain, facts, and SCDs from people who still teach them.

Courses

• Udemy — data modeling / warehouse design

  Useful if the instructor starts with grain, not with a tool logo.

Books

• The Data Warehouse Toolkit, 3rd ed.

  Still the dimensional modeling book. Gold tables are Kimball whether you say it or not.

• Building a Scalable Data Warehouse with Data Vault 2.0

  When you actually need Vault — not as a default for a 20-table warehouse.

YouTube

• Apache Iceberg

  Community talks on catalogs, compaction, and hidden partitioning.

• Tabular / Snowflake Open Catalog sessions

  REST catalog and multi-engine reads — the 2026 interview topic.

Courses

• Iceberg docs (official)

  Start with spec + Spark/Trino quickstarts. Paid courses lag the spec.

Books

• Apache Iceberg: The Definitive Guide (O'Reilly)

  Metadata, catalogs, and maintenance. The book to own if you run a lake.

YouTube

• Dagster

  Assets, checks, and Dagster+ walkthroughs from the makers.

Courses

• Dagster docs (official)

  Start with software-defined assets. Skip random Udemy clones of Airflow DAGs in Dagster syntax.

Books

• Fundamentals of Data Engineering (Reis & Housley)

  Orchestration as a lifecycle choice, not a brand. Pair with the official Dagster tutorial.

YouTube

• Prefect

  Flows, deployments, and work pools from the product team.

Courses

• Prefect docs (official)

  The getting-started path is enough. Paid courses lag the 2.x/3.x APIs.

Books

• Data Pipelines with Apache Airflow (O'Reilly)

  The orchestration mental model still applies. Prefect is a lighter control plane, not a different job.

YouTube

• Apache Flink

  Flink Forward talks — watermarks and state, not slideware.

Courses

• Flink training (official)

  Docs + the training repo. Udemy courses that skip event time are a waste.

Books

• Stream Processing with Apache Flink (O'Reilly)

  The book. Pair with current docs for the Table API and new checkpoint options.

• Streaming Systems (Akidau et al.)

  Watermarks and event time — the model Flink implements.

YouTube

• ClickHouse

  Official meetups on MergeTree, inserts, and query performance.

Courses

• ClickHouse Academy (official)

  Start here. Schema and ORDER BY before cluster topology.

Books

• ClickHouse docs — MergeTree engines

  The engine family is the book. Read this before any paid course.

YouTube

• Trino

  Community talks on connectors, Iceberg, and cost-based optimization.

Courses

• Trino docs (official)

  Catalogs and connectors first. Starburst University if you are on the commercial build.

Books

• Trino: The Definitive Guide (O'Reilly)

  Federation, security, and why SELECT * kills clusters.

YouTube

• Postgres Conference / PGConf talks

  EXPLAIN, vacuum, and logical replication from people who run it.

Courses

• PostgreSQL docs — performance tips

  Official. Pair with EXPLAIN on a real slow query, not a course video.

Books

• PostgreSQL 14 Internals (or current edition)

  How the heap and WAL actually work — the CDC prerequisite.

• Designing Data-Intensive Applications (Kleppmann)

  Replication and logs. Why Debezium exists.

YouTube

• Airbyte

  Connector and platform walkthroughs. Filter for recent major versions.

Courses

• Airbyte docs (official)

  Sources, destinations, and the raw vs normalized decision. That is the course.

Books

• Fundamentals of Data Engineering (Reis & Housley)

  Ingestion as a lifecycle stage. Airbyte is one implementation.

YouTube

• Fivetran

  MAR, destinations, and schema change talks from the vendor. Treat pricing as marketing until you model it.

Courses

• Fivetran docs (official)

  Connector reference and destination write patterns. Then dbt on top.

Books

• The Data Warehouse Toolkit (Kimball)

  Fivetran loads raw. Grain still comes from you.

YouTube

• Debezium community talks

  Snapshots, incremental snapshots, and slot lag — watch recent ones.

Courses

• Debezium docs (official)

  Your database chapter first (Postgres/MySQL). Then the sink MERGE.

Books

• Designing Data-Intensive Applications (Kleppmann)

  Change logs and replication. The theory Debezium implements.

YouTube

• Microsoft Fabric

  Official and MVP sessions. Prefer 2025+ — the item model moved.

Courses

• Microsoft Learn — Fabric

  Official modules before any Udemy “Fabric in 10 hours” clone.

Books

• Microsoft Fabric documentation

  OneLake and capacity. Books lag the product; docs do not.

YouTube

• AWS Events — Redshift

  re:Invent sessions on RA3, Spectrum, and sort keys.

Courses

• AWS Skill Builder — Redshift

  Official. Pair with STL/SVL query practice on a real cluster.

Books

• AWS Redshift Database Developer Guide

  Distribution, sort keys, and WLM. Still the source of truth.

YouTube

• HashiCorp

  State, modules, and provider talks. Skip the 2019 “first resource” videos.

Courses

• HashiCorp Terraform tutorials

  Official. Then apply it to a warehouse + IAM module, not another toy VPC.

Books

• Terraform: Up & Running (O'Reilly)

  State and modules. Pair with the Snowflake or Databricks provider docs.

YouTube

• CNCF / KubeCon data talks

  Spark/Airflow/Flink operators — that is the DE slice, not CKA cram.

Courses

• Kubernetes docs — concepts

  Pods, jobs, requests/limits. Enough to read a crash loop. CKA is optional.

Books

• Kubernetes: Up and Running (O'Reilly)

  The overview. Then read the operator docs for the engine you actually run.

YouTube

• Polars

  Lazy frames and expressions from the maintainers.

Courses

• Polars user guide

  Official. Do the lazy vs eager chapter before rewriting pandas.

Books

• Python for Data Analysis, 3rd ed.

  The DataFrame mental model. Then migrate the slow jobs to Polars.

YouTube

• Great Expectations / GX

  Expectations and suites. Pair with dbt test talks, not only GX marketing.

Courses

• GX docs (official)

  Start with a few expectations on grain. Skip enterprise tours until you have owners.

Books

• Fundamentals of Data Engineering (Reis & Housley)

  Quality as a lifecycle control, not a tool purchase.

YouTube

• Looker / Looker Studio sessions

  LookML and metrics. Prefer talks that start with grain.

Courses

• dbt Semantic Layer / MetricFlow docs

  Metrics as code next to models. Then whatever BI tool your org already paid for.

Books

• The Data Warehouse Toolkit (Kimball)

  If the metric is wrong, it is usually grain — not the chart type.