Senior Data Engineer

Our client is looking for a Senior Data Engineer to design, scale, and maintain high-performance ingestion, Change Data Capture (CDC), and delivery pipelines. In this role, you will architect enterprise-grade data platforms to process large-scale metadata, algorithmic feature extraction, and licensable asset datasets.

This team sits at the foundation of the data processing and feature enrichment pipelines. Driving consistency and performance across an ever-growing data lake, we provide guardrails and sound system design patterns to inform the creation and fine-tuning of the industry's cutting-edge foundational large language models.

You will leverage Google Cloud Platform (GCP) as the primary compute foundation of a multi-cloud data processing ecosystem, alongside hybrid data warehouses and query engines including Snowflake and ClickHouse to provide low-latency data exposure, incremental delta tracking, and seamless integration endpoints for downstream AI/ML consumption.

Primary Tech Stack: GCP, ClickHouse, Snowflake, dbt Core, Apache Airflow, Python, SQL.

Responsibilities

1. Ingestion Architecture & GCP Foundation:

- Architect, build, and optimize scalable job creation, batching, and high-throughput ingestion pipelines hosted on GCP.

- Implement serverless and containerized data pipelines using GCP infrastructure (BigQuery, Dataflow, Pub/Sub, Cloud Storage, GKE).

2. Enterprise ETL, CDC & Incremental Processing:

- Design robust Change Data Capture (CDC) and incremental update workflows to enable fast delta synchronization across platforms.

- Build dbt transformation models and Airflow orchestrations following Data Vault / dimensional modeling patterns.

- Implement delta tracking methods (e.g., hash keys, hash diffs) to track row-level asset modifications and metadata events.

3. Multi-Engine Analytics & Query Optimization:

- Maintain and optimize Snowflake integration layers, handling deduplication logic, canonical content filtering, and policy restrictions.

- Deploy and manage ClickHouse as a high-concurrency analytical engine to deliver sub-second metadata retrieval and indexing.

- Construct dedicated integration views and API query layers allowing internal and external clients to consume real-time delta updates.

4. Data Quality, Governance & Scalability:

- Enforce metadata schema standardization, canonical asset mapping, and automated validation layers across diverse asset types (images, video, 3D models, audio).

- Collaborate closely with cross-functional teams to integrate metadata pipelines with feature extraction tools, search indexes, and licensing platforms.

Requirements

- Google Cloud Platform (GCP): 4+ years of production experience with core GCP data technologies (BigQuery, Dataflow, Cloud Pub/Sub, GKE, Google Cloud Storage).

- Database & Query Engines: Deep technical proficiency with both Snowflake (warehousing, semi-structured queries, integration layers) and ClickHouse (real-time aggregation, low-latency columnar storage, indexing).

- Enterprise ETL / CDC: Proven experience architecting CDC pipelines and incremental delta loading for multi-terabyte dataset scale.

- Data Transformation & Orchestration: Hands-on mastery of dbt Core/Cloud and Apache Airflow in an enterprise CI/CD setting (e.g., GitHub Actions).

- Programming Skills: Advanced skills in Python and SQL; familiarity with Go, Java, or Scala is a strong plus.

Nice-to-Have Qualifications:

- Experience with Data Vault 2.0 architecture, hash-key generation, and raw/staging/integration layer design.

- Background handling AI/ML metadata, computer vision feature vectors, C2PA provenance headers, or semantic search configurations.

- Familiarity with media metadata schemas (images, 3D model stems, video derivatives).

- Experience with other major cloud providers such as AWS, Azure, or Oracle Cloud.

Key Success Metrics (First 90 Days)

- GCP Pipeline Optimization: Standardize ingestion batching and pipeline compute infrastructure on GCP.

- ClickHouse & Snowflake Integration: Establish streamlined integration views exposing low-latency metadata delta changes.

- Automated CDC Workflows: Finalize automated dbt delta transformations with hash tracking across primary warehouse layers.