Role: Sr. Data Engineer with CDP, Google Cloud Platform, Cloud Spanner
Location: Atlanta, GA (hybrid)
Duration: 12+ Months CTH
About the role:
The Data Engineer, CDP builds and operates the real-time pipelines behind Customer Data Platform the system of record for 100M+ customer profiles, now moving to Google Cloud to deliver real-time customer context, near-real-time signals, and sub-50ms API access. Any agent, store associate, or digital channel should see a customer's prior conversation and continue from there.
You will build how clickstream, care, store, and network data flows into Cloud Spanner (low-latency store), BigQuery (analytics), and the knowledge graph behind GraphQL and MCP servers while migrating and reusing proven ETL from the Azure estate (Databricks, Snowflake, Data Factory, Event Hubs, Cosmos DB).
Responsibilities:
- Real-Time Pipeline Engineering
- Build and run NRT ingestion with Pub/Sub and Dataflow (Beam) into Spanner and BigQuery, with data contracts, schema evolution, exactly-once semantics, dead-letter handling, and replay
- Implement Spanner loads and change streams for current-state customer profiles, tuning keys, batching, and hotspots to hold sub-50ms reads
- Clickstream, CDC & Data Modeling
- Implement reusable CDC, SCD Type 2, and late-arriving-data patterns in Beam, SQL, and PySpark; build clickstream ingestion at scale
- Partner with NBA and T-Engage to cut signal latency to sub-five minutes; build BigQuery Spanner sync and reverse ETL
- Azure-to-Google Cloud Platform Migration
- Move data from Azure to Google Cloud Platform with STS and BigQuery Spanner sync, reusing Databricks, Data Factory, and Snowflake ETL; execute migration or retirement of Cosmos DB, Redis, Event Hubs, and Functions with dual-run, reconciliation, and rollback
- Knowledge Graph, API & MCP Data Feeds
- Build the pipelines that populate the customer knowledge graph on Spanner (entity and identity resolution, Customer 360) and feed the sub-50ms GraphQL API and the MCP servers on Cloud Run used by AI agents
- Quality, Operations & Delivery
- Implement data quality, lineage, and access controls (Dataplex, Unity Catalog, TISS-310) and IAM/VPC-SC/CMEK; ship on Cloud Run, Dataproc, Dataflow, and Composer with Terraform, CI/CD, monitoring, SLOs, and cost tuning
- Own pipeline on-call and runbooks, join design reviews, mentor junior engineers, and partner with product, analytics, and AI/ML teams
Requirements:
- Overall (5+ years): Bachelor's in Computer Science or related field (or equivalent); 5+ years in data engineering, 2+ building real-time pipelines in production.
- Google Cloud data platform (3+ years): Production data pipelines on Google Cloud Platform, including IAM, VPC Service Controls, and cost awareness.
- Streaming & NRT pipelines (3+ years): Pub/Sub, Dataflow (Beam), Kafka, Event Hubs, or Spark Structured Streaming with exactly-once semantics, watermarking, and clickstream ingestion at scale.
- BigQuery (3+ years): Partitioning and clustering, reservations and cost control, Spanner federation, and BigQuery-to-Spanner (reverse ETL) data movement.
- Cloud Spanner (1+ year): Schema and primary key design, indexes, hotspot avoidance, change streams, and bulk and streaming loads.
- Data modeling & ETL (4+ years): Dimensional, Data Vault, and medallion models; idempotent ETL/ELT (CDC, SCD Type 2); expert SQL, Python, and advanced PySpark.
- Databricks, Spark & Delta Lake (3+ years): Spark tuning, Delta Lake, Delta Live Tables, and Unity Catalog, to reuse and migrate existing ETL.
- Azure & Snowflake (2+ years): Data Factory, Event Hubs, Cosmos DB, Redis, and Functions, plus Snowflake, to support a hybrid-cloud migration.
- Serving & APIs (1+ year): Building data feeds for REST/gRPC or GraphQL services with p99 targets under 50ms; exposure to MCP, LLM/agent, or Vertex AI data access a plus.
- Platform engineering (3+ years): Terraform/IaC, CI/CD, Cloud Run, Dataproc, and Composer or Airflow; observability and SLOs.
- Testing & data quality (2+ years): Unit, integration, and data-quality testing for pipelines; reconciliation, lineage, and Dataplex or similar tools.
- Collaboration (2+ years): Working with product, analytics, and AI/ML teams; clearly explaining design tradeoffs; mentoring junior engineers.
Must-Have Skills
- Google Cloud Real-Time Pipelines & Serving
- Real-Time Pipelines Pub/Sub and Dataflow (Beam) exactly-once, watermarking, ordering, dead-letter handling, schema evolution, and replay
- Cloud Spanner Loading: Key design, batching, hotspot avoidance, change streams, and stale reads at 100M+ profile scale
- BigQuery: Partitioning and clustering, reservations and cost governance, BigLake, Spanner federation, and reverse ETL to Spanner
- Graph & API Feeds: Pipelines for identity resolution and Customer 360 on Spanner Graph, feeding GraphQL (sub-50ms p99) and MCP servers on Cloud Run
- Streaming, Analytics & Compute
- Beam & Spark Streaming Apache Beam (Python/Java), Spark Structured Streaming, and Kafka or Event Hubs stateful processing, windowing, and backpressure
- Orchestration & Compute: Composer/Airflow, Cloud Run, Dataproc, Dataflow, GKE, and Cloud Functions service selection and right-sizing
- Pipeline Code Quality: Production Python, Java, SQL, and PySpark unit and integration testing, code review, and performance tuning
- Hybrid Cloud & Existing Azure Platform
- Cross-Cloud Data Movement: STS, BigQuery Data Transfer, Interconnect/ExpressRoute, Workload Identity Federation, egress cost, and Azure-to-Google Cloud Platform reconciliation
- Databricks & Delta Lake: Databricks, Delta Lake, DLT, Unity Catalog, and medallion architecture to reuse and migrate existing ETL
- Snowflake & Azure Services: Snowflake; Data Factory, Event Hubs, Cosmos DB, Redis, and Functions to map, replace, or retire
- Modeling, Governance & Operations
- Data Modeling & ETL: Dimensional, Data Vault, and medallion models; idempotent ETL/ELT (CDC, SCD Type 2, late-arriving data); expert SQL and PySpark
- Security, Quality & Ops: IAM, VPC-SC, CMEK, Cloud DLP, Dataplex, Unity Catalog, TISS-310 alignment; monitoring, SLOs, Terraform, CI/CD, and cost tuning
In compliance with the salary transparency law, the expected pay range for this role is $65/hr. Actual compensation depends on experience and interview evaluation.
