Data Engineer
About the role
Our client is an innovative, fast-growing AI startup transforming the marketing technology landscape and is looking to add a Data Engineer to its engineering team. This is an opportunity to make a significant impact by building the core data platform that powers the company's AI-driven products. In this role, you'll own the end-to-end data pipeline, from ingesting and transforming large volumes of marketing data to designing warehouse models, defining the metrics that drive the business, and building the AI insights engine that converts raw data into intelligent, actionable recommendations.
Responsibilities Own and evolve the warehouse transformation models, from raw external tables and incremental staging, through conformed marts (period-over-period, rolling, and cumulative-to-date), to the insights layer. Own the ingestion of marketing data: connection provisioning for new source accounts, plus the storage landing to raw to warehouse flow. Build and harden data quality, freshness, and schema-change handling so a vendor changing a field upstream never silently breaks a brand’s reporting. Own the metric registry and grain definitions that the data API compiles into warehouse SQL, the canonical definitions of every metric the product reports. Build out the AI-insights and recommendation pipeline: author and tune the insight detectors (creative fatigue, spend inefficiency, channel saturation, pacing, and more), do the prompt and evaluation work against LLM APIs, and make recommended-action generation trustworthy. Guarantee multi-tenant correctness across every model, with isolation enforced at the data layer and via row-level security, so numbers never cross tenant boundaries. Tie into the nightly orchestration pipeline (ingest, transform, dimension sync, insight generation, cache flush) and keep insight quality high as data volume grows. Instrument the pipeline with error monitoring and structured logs so you can debug a bad number or a stale insight quickly, and feed into SOC 2 and ISO 27001 readiness as data controls mature.
Requirements Strong SQL and real experience with a layered, modular warehouse transformation project. Solid Python for data work across ingestion, transformation glue, and pipeline tooling. Genuine analytical rigor: you care whether a metric is correct, reason about grain and aggregation, and have debugged why a number is wrong all the way to root cause. Comfort working in a cloud data warehouse and reasoning about partitioning, clustering, and query cost. Hands-on with LLM APIs for a real product feature across prompting, evaluating outputs, and handling nondeterminism, or a strong appetite and aptitude to own it. High ownership and comfort with early-stage ambiguity; you can take a fuzzy goal and turn it into shipped detectors and metrics. Experience within the marketing industry and marketing data Familiarity with EL/ELT ingestion tooling is nice to have Exposure to multi-tenant data systems and row-level-security or tenant-isolation patterns. Experience building LLM evaluations or detector and recommendation systems with measurable quality bars. Interest in or experience with SOC 2 or ISO 27001 data controls and governance.
Similar Jobs
Data Engineer
About the role
Our client is an innovative, fast-growing AI startup transforming the marketing technology landscape and is looking to add a Data Engineer to its engineering team. This is an opportunity to make a significant impact by building the core data platform that powers the company's AI-driven products. In this role, you'll own the end-to-end data pipeline, from ingesting and transforming large volumes of marketing data to designing warehouse models, defining the metrics that drive the business, and building the AI insights engine that converts raw data into intelligent, actionable recommendations.
Responsibilities Own and evolve the warehouse transformation models, from raw external tables and incremental staging, through conformed marts (period-over-period, rolling, and cumulative-to-date), to the insights layer. Own the ingestion of marketing data: connection provisioning for new source accounts, plus the storage landing to raw to warehouse flow. Build and harden data quality, freshness, and schema-change handling so a vendor changing a field upstream never silently breaks a brand’s reporting. Own the metric registry and grain definitions that the data API compiles into warehouse SQL, the canonical definitions of every metric the product reports. Build out the AI-insights and recommendation pipeline: author and tune the insight detectors (creative fatigue, spend inefficiency, channel saturation, pacing, and more), do the prompt and evaluation work against LLM APIs, and make recommended-action generation trustworthy. Guarantee multi-tenant correctness across every model, with isolation enforced at the data layer and via row-level security, so numbers never cross tenant boundaries. Tie into the nightly orchestration pipeline (ingest, transform, dimension sync, insight generation, cache flush) and keep insight quality high as data volume grows. Instrument the pipeline with error monitoring and structured logs so you can debug a bad number or a stale insight quickly, and feed into SOC 2 and ISO 27001 readiness as data controls mature.
Requirements Strong SQL and real experience with a layered, modular warehouse transformation project. Solid Python for data work across ingestion, transformation glue, and pipeline tooling. Genuine analytical rigor: you care whether a metric is correct, reason about grain and aggregation, and have debugged why a number is wrong all the way to root cause. Comfort working in a cloud data warehouse and reasoning about partitioning, clustering, and query cost. Hands-on with LLM APIs for a real product feature across prompting, evaluating outputs, and handling nondeterminism, or a strong appetite and aptitude to own it. High ownership and comfort with early-stage ambiguity; you can take a fuzzy goal and turn it into shipped detectors and metrics. Experience within the marketing industry and marketing data Familiarity with EL/ELT ingestion tooling is nice to have Exposure to multi-tenant data systems and row-level-security or tenant-isolation patterns. Experience building LLM evaluations or detector and recommendation systems with measurable quality bars. Interest in or experience with SOC 2 or ISO 27001 data controls and governance.