Senior Data Engineer

We are building a central Data Engineering team responsible for ingesting data from enterprise source systems and engineering applications — including Oracle, CRM platforms, and internal engineering tools — loading it into our data platform and delivering foundational data products that serve Group Practices across the firm. We're looking for a senior engineer who combines an architecture mindset with deep hands-on experience — someone who can design scalable systems and personally build and ship them into production. You will own pipelines end-to-end, from whiteboard to running code, on Google Cloud.

What You'll Do

ARCHITECTURE & DESIGN

• Architect end-to-end data platform solutions: from source system connectivity through ingestion, storage, transformation, and data product delivery — making deliberate decisions on scalability, cost, and maintainability.

• Define reference architectures and reusable patterns: for pipelines, storage layout, IAM boundaries, and CIM modeling that the whole engineering team builds from.

• Evaluate and select Google Cloud services: for each workload: streaming vs. batch, event-driven vs. scheduled, structured vs. unstructured.

• Translate business requirements: from Portfolio Data Managers and domain data owners into clear, implementable technical designs.

HANDS-ON PIPELINE & PRODUCTION ENGINEERING

• Personally build and ship production pipelines: this role codes, not just designs. You will own your work from first commit through deployment, monitoring, and ongoing reliability.

• Build end-to-end ingestion pipelines: from source systems (Oracle, CRM, engineering applications) into Google Cloud Storage and BigQuery.

• Build multimodal unstructured data processing workflows: handling PDFs, Excel, PowerPoint, audio recordings, video files, and engineering drawings.

• Apply Google Cloud AI services: Document AI, Speech-to-Text, Video Intelligence, Vision API, Vertex AI — to extract structured information from unstructured content and ship results to production datasets.

• Use LLMs: to classify, summarize, and transform unstructured text and documents into structured, queryable data products.

• Build and maintain Common Information Model (CIM) semantic layers: aligning ingested data with business-domain meaning across Group Practices.

• Design and manage Google Cloud Storage buckets: lifecycle policies, folder structure, access controls, and cost optimization.

Required Skills

ARCHITECTURE MINDSET & PRODUCTION ENGINEERING

• Proven record of designing systems and then personally building them for production, not delegating implementation after the whiteboard session.

• Architecture trade-off decisions: streaming vs. batch, managed vs. custom, normalized vs. denormalized — and the ability to explain the reasoning clearly.

• Reliability patterns: idempotency, dead-letter queues, retry logic, schema evolution, and data quality gates.

• Technical design documents that a team can build from, and are equally comfortable writing the code.

GOOGLE CLOUD PLATFORM

BigQuery · Cloud Storage (GCS) · Cloud IAM · Dataflow / Apache Beam · Cloud Composer / Airflow · Pub/Sub · Vertex AI · Cloud Functions / Run

UNSTRUCTURED DATA & AI TRANSFORMATION

Google Document AI · Speech-to-Text API · Video Intelligence API · Vision API / OCR · PDF & Office file parsing · Audio processing · Engineering drawing extraction · LLM-based data extraction

DATA ENGINEERING FUNDAMENTALS

Python · SQL (BigQuery dialect) · dbt · ELT / ETL pipeline design · Data modeling · Git / CI-CD · Terraform / IaC

SECURITY & GOVERNANCE

• Google Cloud IAM: service accounts, custom roles, workload identity federation, least-privilege design across projects and datasets.

• GCS access controls: uniform bucket-level access, IAM conditions, VPC Service Controls, audit logging.

DATA MODELING & CIM

• Common Information Model (CIM): experience building CIM semantic layers — ideally in energy, utilities, or infrastructure domains (IEC 61968 / 61970).

• Source-aligned data products: faithful to source systems, enriched with business-domain meaning ("zero layer").

• Dimensional modeling: medallion architecture (bronze / silver / gold), and structuring data for downstream BI and AI consumers.