Data Engineer, AI & Analytics

About the role

Power Digital is seeking a Data Engineer to join its Data Team, responsible for building and maintaining the core data foundation that powers agency, client and AI initiatives. The role involves end‑to‑end work from raw marketing data ingestion through semantic layers, leveraging AI‑agentic workflows to accelerate development.

Key responsibilities

• Design, build, and maintain ingestion pipelines, data models and data marts for multi‑client marketing data.
• Handle API changes, rate limits and retroactive conversion restatements without corrupting downstream models.
• Model spend, impressions, conversions and revenue across Meta, Google, TikTok, Amazon, LinkedIn, Microsoft and integrate customer‑level data from Shopify, Klaviyo, GA4 and CRMs.
• Develop client‑specific modeling on top of the core layer while preserving a shared foundation.
• Create semantic layers and metric definitions for AI‑generated SQL queries.
• Use AI coding tools (e.g., Cursor, Claude Code, GitHub Copilot) to speed development and document reusable patterns.
• Collaborate with product, engineering, AI/innovation and client teams to translate requirements into data solutions.
• Monitor data quality, resolve issues and optimise pipelines for cost and performance in a multi‑tenant warehouse.

Required profile

• 3+ years of experience in data or analytics engineering, including ownership of a production dbt project.
• Advanced proficiency in Python and SQL for production‑grade data pipelines.
• Deep expertise with dbt (incremental strategies, Jinja, macros, tests, snapshots, exposures).
• Strong command of Snowflake and the broader cloud data stack.
• Experience modelling in multi‑tenant environments and handling marketing/advertising datasets.
• Familiarity with GCP and infrastructure‑as‑code principles.
• Proven use of AI‑agentic development workflows and CI/CD best practices.

Required skills

• Python
• SQL
• dbt
• Snowflake
• Google Cloud Platform (GCP)
• Git
• CI/CD pipelines
• AI coding assistants (Cursor, Claude Code, GitHub Copilot)
• Feature store design and semantic layer modeling