Director, Analytics

Cognitiv•Bellevue, WA

1d•$210,000 - $230,000•Hybrid

About The Position

Are you ready to revolutionize the advertising industry? At Cognitiv, we are not just another AdTech company—we are industry trailblazers redefining media buying with our Deep Learning Advertising Platform. Since 2015, we have harnessed the power of cutting-edge deep learning technology and data science to transform how brands connect with their customers. Our mission? To bring intelligence to advertising and deliver unparalleled precision, relevance, and impact at scale. With our innovative platform, advertisers enjoy unprecedented flexibility—whether it is activating Dynamic Deals through their preferred DSP, leveraging our managed service DSP, or utilizing our industry-first ContextGPT product. As a part of Cognitiv, you will be at the forefront of AI-driven advertising solutions, driving change and achieving remarkable growth in a rapidly evolving industry. Now, we’re growing! The Mission Own the real-time performance control layer that turns model output into revenue, margin, and delivery stability. Your systems will prevent spend-outs, enforce margin floors, surface anomalies early, and convert prediction quality into collected dollars—at bid-time and at scale. Why This Role Is Exciting Direct P&L impact: Your work controls real dollars—take-rate, margin, and delivery. Real-time scale: Billions of bid requests per day with hard latency and reliability constraints. High leverage scope: The control layer every campaign depends on. Build > maintain: Greenfield upgrades to autotuners, SLOs, dashboards, and incident processes. Cross-functional seat: Influence Science guardrails, Trading execution, and Platform roadmap. Career runway: Executive visibility with wins tied to ARR and take-rate.

Requirements

A seasoned AdTech operator. You bring 8–12+ years in AdTech or performance platforms, with 3–5+ years leading DS or analytics teams in production environments.
Fluent in real-time systems. You have deep, hands-on knowledge of RTB, pacing, budget control, and bid-time decisioning—and understand how small changes ripple at scale.
A builder of control loops, not dashboards. You’ve designed closed-loop systems where monitoring leads to action, SLOs are real, and reliability is measured—not assumed.
Causality-minded and decision-oriented. You know how to design experiments, interpret results, and turn statistical insight into product and operational decisions.
Hands-on and technically credible. You’re expert in SQL and Python and comfortable debugging production data pipelines, logs, and metrics when things break.
An outcomes-first leader. You think like an operator: reliability, margin, and delivery matter more than elegant theory—and you hold yourself accountable to real-world impact.

Nice To Haves

Experience with identity graphs, cross-device attribution, CTV, viewability, or IVT.
Familiarity with online inference systems, feature stores, or latency optimization.
Exposure to MMM or econometrics and how it complements incrementality testing.
Prior ownership of incident management, error budgets, or on-call processes.

Responsibilities

Own the performance control layer. Lead a DS/analytics team responsible for autotuners, pacing systems, and real-time campaign health—turning model output into stable delivery and margin.
Build reliability into revenue. Define and own performance and reliability SLOs, alerting, and a single-pane dashboard that makes pacing and margin issues obvious before they impact spend.
Kill discrepancies at the root. Run deep forensics across clicks, impressions, and conversions to identify root causes and ship durable fixes—not one-off workarounds.
Prove what actually drives lift. Design and run incrementality experiments (A/B, PSA control, geo-lift), interpret results, and publish clear, client-ready readouts that inform decisions.
Translate models into operations. Partner with Science to set constraints and guardrails, and with Trading to operationalize changes safely in live environments.
Set the operating standard. Establish incident response, postmortems, and playbooks that enable fast recovery, clear ownership, and continuous improvement.