Sr Data Engineer (Data Architecture and Modeling)

Paramount•New York, NY

34d•$110,000 - $165,000

About The Position

The Signal Intelligence (SigInt) team turns a high-volume, multi-source data ecosystem — operational systems, client-app event streams, ad-tech platforms, stitcher events, ad server logs, and telemetry — into the trusted data models, metrics, and dashboards that drive business, product, and ad-ops decisions. We are hiring a Senior Data Engineer who is, in equal measure, a data modeler, a SQL craftsperson, and an analyst — not only a pipeline builder. You will own data end-to-end: modeling raw events into conformed, analytics-ready marts; writing the analytical SQL that answers hard business questions; defining the metrics leadership relies on; and building the pipelines that feed it all. Because we sit close to the business, translating ambiguous questions into durable models and clear insight matters as much as engineering the pipeline. In practice, you will be the team's data sherpa — the person who knows the data terrain cold, guides analysts and business partners through it, and turns an ambiguous “can we find out X?” into a trusted, well-modeled answer. You own the map — the definitions, lineage, and the “why” behind the numbers — as much as the pipelines. This is an analytics-engineering role on a semi-business team. It is not a pure platform / infrastructure position — if your passion is exclusively distributed-systems plumbing with little interest in modeling, analysis, or guiding the business, this will not be the right fit.

Requirements

Expert SQL, including analytical SQL craft (window functions, complex joins, cohort / funnel / sessionization logic) and performance tuning on large event-scale datasets.
Deep data-modeling expertise — dimensional modeling, star/snowflake schemas, slowly changing dimensions, conformed dimensions, grain, and data-mart design; able to design a trusted source of truth from messy multi-source data.
Strong data analysis and business translation — able to independently investigate data, define metrics and KPIs, and act as a trusted guide who turns ambiguous business questions into clear answers and enables others to self-serve.
A guide’s temperament — curiosity to learn the data terrain deeply, patience to help non-technical partners, and communication that earns trust across the business.
dbt (or an equivalent transformation / modeling framework).
Cloud warehouse — BigQuery strongly preferred (our environment); Snowflake or Databricks acceptable with a fast ramp.
Python for data processing and pipeline development.
Orchestration experience (Airflow).
High-volume event data — clickstream, telemetry, ad impressions, or similar — including deduplication, late-arriving data, sessionization, and schema evolution.
Proven track record harmonizing data across multiple source systems with conflicting schemas, identifiers, or grain.
BI tools (DOMO, Looker, Mode) — comfort both consuming and supporting their development.
Experience debugging data-quality issues across the full stack — from BI tool to warehouse to raw event logs.
BA/BS in Computer Science, Math, Physics, Engineering, Economics, Statistics, or a related quantitative field.
Strong analytical, logical, and communication skills.

Nice To Haves

Semantic-layer experience (LookML, dbt Semantic Layer, Cube).
Streaming-media / ad-tech domain (ad servers, SSAI / stitchers, VAST/VMAP, impression and fill-rate metrics).
Streaming ingestion (Kafka, Pub/Sub, Kinesis).
Statistical analysis with R and predictive-analytics tooling; experimentation / A-B analysis.
Master data management or entity resolution.
MS in a quantitative discipline; experience mentoring or leading engineers/analysts.

Responsibilities

Design and own the dimensional and semantic models — conformed dimensions, facts, star/snowflake schemas, slowly changing dimensions (Type 2), and well-defined grain — that serve as the trusted source of truth.
Architect data marts spanning operational, advertising, and telemetry domains for analytics, reporting, AI, and operational use.
Establish shared definitions and a semantic / metrics layer so users, sessions, devices, content, campaigns, and impressions mean the same thing everywhere.
Write and optimize sophisticated analytical SQL — window functions, cohort and funnel logic, sessionization, distributions — on large event-scale datasets.
Analyze complex datasets to find patterns, quantify drivers, and produce actionable insight; define new metrics and KPIs (usage, revenue, etc.) and the logic behind them.
Independently investigate variance across platforms and communicate findings clearly to peers and senior leadership.
Be the go-to guide for the data — the person who knows where every number comes from, what it means, and its quirks and landmines.
Own the institutional knowledge of the data — definitions, lineage, and the reasoning behind metrics — and make it legible so analysts, ad ops, product, and finance can self-serve with confidence.
Partner with stakeholders and source-system owners to translate ambiguous business questions into durable data models and metrics, and support/conduct strategic investigations and analysis.
Build and harmonize the ingestion and transformation pipelines — event data from client apps, ad-tech platforms, stitchers, ad servers, and telemetry — that feed the models, handling deduplication, late-arriving data, sessionization, and schema evolution.
Stitch identity and sessions across client, server, and ad-side events for accurate user, content, and revenue analytics.
Establish data-quality standards (testing, monitoring, alerting, freshness and volume SLAs) and troubleshoot incidents end-to-end — from a dashboard anomaly back through marts, transformations, and raw logs.
Support and improve BI dashboards; document datasets, lineage, and business logic so the data map stays trustworthy and current.
Mentor engineers and analysts on SQL, modeling, event data, and engineering best practices.