Venatus - Analytics Engineer / Data Scientist (Mid–Senior)

• At Venatus, we pride ourselves on having the most detailed and robust data available in the industry for every campaign type - whether it be a programmatic auction or a direct sale. This role owns the step-change from “data exists” to “data is consumable”: trusted, well-modelled, well-documented, and easily usable across the business. You will do this through creating dashboards, self-serve semantic metrics, and next-gen agentic experiences. • You will sit at the intersection of Data Engineering, Analytics, and the business, shaping how the company interacts with data day-to-day. • What you’ll own • Build the company’s self-service data product • Design and deliver clean, scalable, well-documented datasets (marts / data products) that become the default “source of truth”. • Define a practical consumption strategy: what belongs in dashboards, what belongs in curated datasets, and what belongs in self-serve exploration and agentic workflows. • Tooling and architecture leadership • Lead the evaluation and selection of tooling for modelling, semantic metrics, catalog/dictionary, and consumption (dashboarding and/or agentic interfaces). • Partner with Data Engineering to align on reference architecture, layering, SLAs, environments, CI/CD, and cost/performance trade-offs. • Own the semantic layer & metrics governance • Build and maintain the semantic layer (canonical metrics, dimensions, metric definitions, and reusable logic). • Create and curate the data dictionary / business glossary: naming conventions, definitions, lineage, owners, and change management. • Establish “definition of done” standards: tests, documentation, review, and release discipline. Own the verification and validation of all data outputs. • Make data ML-ready • Shape datasets for machine learning consumption: feature-ready tables, consistent entity definitions, time semantics, training/serving considerations, and backfill strategies. • Collaborate with Data Science / ML engineering to ensure models are supported by stable, observable upstream data products. • Enable streaming and near-real-time use cases • Work with Data Engineering to enable streaming/near-real-time datasets where they materially improve decisions (e.g., operational monitoring, pacing, anomaly detection). • Define modelling patterns for incremental/streaming data (late arrivals, dedupe, idempotency, watermarking, and quality checks). • Drive adoption through enablement • Train the business in how to use the semantic layer, self-serve datasets, and tooling: onboarding, documentation, office hours, and internal playbooks. • Become the “bridge”: convert stakeholder questions into measurable metrics and scalable data products. • What success looks like: • You’ve mapped the top consumption journeys (exec reporting, revenue, publisher, finance) and identified the top “trust blockers”. • You’ve proposed a target architecture for modelling + semantic layer + dictionary (and the migration path). • First golden data products shipped with tests + documentation + owners. • Semantic layer MVP live for a core domain, with visible adoption (usage + reduced ad hoc). • Self-serve has meaningfully reduced manual reporting burden; key metrics are consistently defined. • ML teams have stable, feature-ready datasets with clear lineage and SLAs. • Streaming/near-real-time datasets exist where they matter, with observability and quality gates. • Expert SQL and strong data modelling fundamentals (dimensional modelling, incremental patterns, performance tuning). Working knowledge of Python and experience with notebooks. • Hands-on experience with a modern modelling workflow (e.g., dbt or equivalent) and strong Git + review discipline. • Practical experience establishing semantic metrics, definitions, and governance that stick (not a theoretical catalog nobody uses). • Strong stakeholder capability: you can translate ambiguous business needs into clear, measurable, testable data products. • You can operate as a technical lead: set standards, unblock others, and drive decisions. • Familiarity with ad-tech / retail media concepts: impressions, fill, pacing, yield, attribution/measurement, identity/privacy constraints. • Experience with data observability, FinOps/cost governance, and production-grade data quality practices. • Experience building data products for activation/audiences/measurement use cases or privacy-aware data collaboration patterns. • Typical tools you'll use: • Warehouses/lakes: Clickhouse • Transformation: dbt • Orchestration: Dagster