SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Zefir is building an AI autopilot for home sales across Europe, starting in France. An AI agent orchestrates the entire property transaction end-to-end, coordinating local brokers, portals, buyers, and documents. Backed by $55M+ from top-tier investors including Sequoia Capital, the company is scaling rapidly.
This is the first dedicated data platform hire in years. The foundations—ingestion and centralization into BigQuery—already work, but everything downstream is missing: canonical data models, unified metric definitions, cost control, and self-serve analytics layers. Today the stack holds because individuals across Operations, Growth, Finance, and Engineering compensate locally with workarounds. KPIs drift between tools, tracking breaks silently, costs escalate, and every question routes through one person.
You will own the entire data platform layer as the single accountable owner. This is not a support function or BI factory—it is the semantic and governance foundation that enables the entire organization to make decisions on trustworthy data.
Key responsibilities include: designing and documenting canonical data models with Bronze/Silver/Gold layers and clear entity definitions (Buyer, Seller, Asset, Agent); building a self-serve semantic layer with row- and column-level security so Finance, Growth, Ops, and Account Managers build their own dashboards; establishing global event taxonomy and tracking governance with hybrid client-side and server-side strategies, GDPR-compliant consent flows, and trustworthy attribution; and setting platform reliability standards including tested transformations, freshness/failure monitoring, cost controls, and safe data access patterns for AI agents querying the warehouse directly.
Success after 12 months means: one documented event taxonomy used across Engineering, Growth, and CRM; a semantic layer where every shared KPI has a single definition, owner, and version history; published SLAs and explicit ingestion topology with no production dependencies on BI tables; self-serve dashboards built by Ops, Growth, Finance, and AMs; and trustworthy attribution enabling defensible acquisition spend decisions.
This role suits someone with 7+ years as a Data, Analytics, or Platform Engineer (ideally at fast-moving consumer or marketplace companies with Staff/Lead exposure). Required: hands-on expertise with BigQuery, dbt, advanced SQL, Python pipelines, orchestration tools (Airflow, Dagster, Prefect); production experience shipping event tracking end-to-end; familiarity with ingestion patterns (Fivetran, Airbyte, CDC), reverse-ETL (Hightouch, Census, Segment), and access governance; experience building and owning semantic/metrics layers; treating AI agents as first-class data consumers; strong ownership and communication skills; and fluency in English and French.
There is no data team to manage initially—you build alone before potentially building a team. This offers real autonomy and direct founder access in exchange for solo execution. It suits someone who has led teams and wants to return to hands-on building.
