Data Solutions Architect learning path: everything to study, topic by topic

Star schema, snowflake schema, fact and dimension tables, grain definition, and Kimball methodology

Normal forms (1NF-3NF, BCNF), denormalization strategies, and normalization vs performance trade-offs

Schema evolution, backward/forward compatibility, schema-on-read vs schema-on-write, and NoSQL schema patterns

SCD Types 1-6, hybrid approaches, temporal data modeling, and bitemporal design

ETL vs ELT architectures, transformation strategies, data validation, and pipeline design patterns

Airflow, Dagster, Prefect, Step Functions — DAG design, scheduling, retry policies, and dependency management

Producer-consumer contracts, schema enforcement, SLA definitions, and breaking change management for data interfaces

Designing pipelines that produce consistent results on re-execution: deduplication, upsert patterns, and watermark-based processing

ML pipeline orchestration, feature stores, model registry, model serving infrastructure, monitoring, A/B testing, and data/model versioning for production ML systems

Feature stores for ML, training data pipeline design, model serving data needs, embedding pipelines, and data versioning for reproducibility

Legacy to modern platform migration, zero-downtime migration strategies, dual-write patterns, data validation during migration, and rollback planning

Lakehouse design combining data lake flexibility with warehouse reliability: Delta Lake, Iceberg, Hudi, and medallion architecture

Cloud warehouse architecture: Snowflake, BigQuery, Redshift design patterns, workload management, and query optimization

Table partitioning, data layout optimization, partition pruning, and bucketing for query performance

Parquet, ORC, Avro, JSON, CSV — choosing formats based on read/write patterns, compression, and schema evolution needs

Apache Iceberg, Delta Lake, and Apache Hudi — table format comparison, schema evolution, time travel, partition evolution, and interoperability across engines

Vector databases for AI workloads, graph analytics platforms, serverless data processing, real-time OLAP engines, and evaluating emerging data technologies

Column-level lineage tracking, impact analysis, lineage visualization, and automated lineage extraction from SQL and code

Metadata management, data discovery, business glossary, and catalog governance for large-scale data platforms

Data quality dimensions, automated testing, Great Expectations, dbt tests, and quality SLAs for data products

PII detection, tokenization, masking, anonymization, GDPR right-to-be-forgotten, and privacy-preserving analytics

Implementing data contracts between producers and consumers with schema versioning and backward/forward compatibility

Building end-to-end data observability: freshness, volume, schema, distribution monitoring, and incident response

Encryption at rest and in transit, tokenization strategies, data classification frameworks, access control models, and security architecture for data platforms

Data SLOs and SLAs, incident management for data systems, on-call for data platforms, error budgets, and reliability patterns for data pipelines

Topic design, partitioning strategies, consumer groups, retention policies, and operational patterns for event streaming

Flink, Spark Structured Streaming, ksqlDB — windowing, watermarks, state management, and late data handling

Debezium, DMS, log-based CDC, trigger-based CDC, and CDC integration with streaming platforms

Stream-table duality, incremental materialization, serving layer design, and real-time aggregation patterns

Achieving exactly-once processing in distributed streaming: transactional producers, idempotent consumers, and end-to-end guarantees

Streaming-first architecture design, Kappa vs Lambda architecture, event-driven analytics, real-time decisioning systems, and low-latency data serving

Domain ownership, data as a product, self-serve data platform, and federated computational governance

Data product design, APIs and interfaces, discoverability, SLAs, and lifecycle management for data-as-a-product

Data platform cost optimization: storage tiering, compute right-sizing, query cost attribution, and FinOps for data

Evaluating data platform vendors: Snowflake vs Databricks vs BigQuery, build vs buy, and total cost of ownership analysis

Data maturity models, organizational readiness assessment, center of excellence design, and data literacy programs

LLM integration in the data stack, embedding pipelines, vector databases, RAG for analytics, and AI-assisted data quality for modern data platforms

Designing enterprise-scale data platforms: ingestion, storage tiers, compute, governance, and consumption layers

Evaluating and selecting cloud data warehouses at scale: Snowflake vs BigQuery vs Redshift, cost models, and migration strategies

Designing end-to-end ML platform architecture: data pipelines, feature stores, training, serving, monitoring, and governance

TCO modeling for data platforms, build vs buy decision frameworks, vendor comparison methodologies, hidden cost identification, and long-term cost projection

Data portability across cloud providers, avoiding vendor lock-in, cross-cloud replication, multi-cloud data architectures, and abstraction layer design

Data mesh implementation patterns, domain ownership models, federated governance, data product interoperability, and organizational design for federation

Self-service tooling for data teams, SDK design, data platform as product, developer portals, and internal platform adoption strategies