Snowflake & dbt Masterclass — 100 Hours of Analytics Engineering End to End

• Snowflake SQL specifics: QUALIFY, LATERAL FLATTEN, semi-structured VARIANT, ASOF joins
• Window functions, recursive CTEs, pivots, time-series gap filling
• Dimensional modelling: star schema, SCD types, surrogate keys, fact grain, bridge tables
• Data Vault 2.0 overview and when it earns its complexity
• Lab: model a messy source system into a defensible star schema

• Snowpipe (auto-ingest) and Snowpipe Streaming; Kafka connector
• Streams (standard, append-only, insert-only) and Tasks (serverless, DAGs)
• Dynamic Tables: declarative pipelines, lag targets, refresh modes
• External tables, Iceberg tables, external volumes
• Fivetran/Airbyte ELT patterns and their cost model
• Lab: Project 2 near-real-time pipeline

• Query profile reading: spilling, pruning ratio, exploding joins, remote IO
• Clustering keys, automatic clustering, search optimisation service, materialised views
• Caching layers: result cache, warehouse cache, metadata — and how to measure hit rates
• Warehouse strategy: separation by workload, queuing, multi-cluster economics
• Resource monitors, budgets, ACCOUNT_USAGE queries, per-team chargeback
• Lab: Project 3 cost audit with a written savings recommendation

• RBAC design patterns: functional vs access roles; least privilege at scale
• Dynamic data masking, row access policies, tag-based policies, object tagging
• Network policies, SSO/SCIM, key-pair auth, secrets handling
• Time Travel, Fail-safe, cloning, replication and failover
• Secure data sharing, reader accounts, Snowflake Marketplace, clean rooms
• Lab: GDPR-style masking implementation with an access-test matrix

• Snowpark DataFrame API in Python; pushdown execution model
• UDFs, UDTFs, stored procedures; Python packages and Anaconda channel
• Snowpark Container Services overview; Streamlit in Snowflake
• Cortex AI functions and vector search — LLM features inside the warehouse
• Lab: replace a Spark job with Snowpark and compare cost and runtime

• Project structure, profiles, targets; dbt Core vs dbt Cloud vs dbt Fusion
• Models, ref(), source(), materialisations (view, table, incremental, ephemeral)
• Staging → intermediate → marts layering; naming and folder conventions
• Tests: generic, singular, dbt_utils, dbt_expectations; severity and thresholds
• Documentation, exposures, the docs site, lineage graph
• Lab: build the staging and marts layers of Project 1

• Incremental models: strategies (merge, insert_overwrite, microbatch), is_incremental, late data
• Snapshots for SCD2; seeds; hooks; custom schemas and aliases
• Jinja and macros; packages; custom generic tests; dispatch and adapters
• Semantic layer and MetricFlow; metrics definitions
• Performance: model materialisation choices, DAG shape, threads, warehouse per model
• Lab: convert a 40-minute full refresh into a 3-minute incremental build

• dbt Cloud jobs vs Airflow (Cosmos) vs Dagster — choosing an orchestrator
• Slim CI: state:modified, deferral, PR-scoped builds, blue/green deployments
• GitHub Actions pipeline for dbt: lint (sqlfluff), compile, test, deploy
• Data quality and observability: freshness, anomaly detection, alerting, data contracts
• Lab: full PR-to-production dbt workflow with tests as gates

• Snowflake vs Databricks vs BigQuery vs Redshift — an honest comparison you can defend
• Migration playbooks: Teradata/Oracle/SQL Server → Snowflake
• Capstone presentation and architecture review
• SnowPro Core exam guide, 120 practice questions, two timed mocks
• Analytics engineer interview questions and resume framing

No. The 30-day trial includes $400 of credits, which comfortably covers the course if you keep warehouses on XS with 60-second auto-suspend — taught in Module 1. You can start a fresh trial if you need more time.

You will learn dbt Core (free, open source) as the foundation, and use dbt Cloud's free developer tier for the scheduling and CI features. Everything transfers between them.

The dbt half is directly transferable — dbt runs on Databricks too. The Snowflake half is platform-specific but the modelling, testing and cost-governance thinking carries over. Many students take this alongside the Databricks course.

The course maps to SnowPro Core, with 120 practice questions and two mocks. Advanced Data Engineer topics are touched on but that exam needs additional dedicated study.