Staff Software Engineer, Ads Measurement Signal

Pinterest Pinterest · Consumer · Seattle, WA · Monetization

Staff Software Engineer to lead the backend architecture and implementation of a GenAI-powered Conversion Health agent. The role involves designing and implementing services, data flows, and retrieval layers for agents to reason over conversion data, evolving existing GenAI applications with prompt hardening, retrieval quality improvements, and evaluation/safety checks. It also includes designing connections with downstream ads products, structuring the measurement context layer into AI-consumable signals, and partnering with various teams to translate measurement pain points into agent skills. The role emphasizes leading an evolving GenAI culture, navigating ambiguity, codifying guardrails, and driving engineering velocity and measurement quality.

What you'd actually do

  1. Own the design and implementation of the Conversion GenAI agent: services, data flows, and retrieval layers that let agents reason over EQS, conversion funnels, and their impact on performance at scale.
  2. Evolve the existing GenAI Applications: harden prompts and tools, improve retrieval quality, add evaluation and safety checks, and make the agent reliable enough for always-on monitoring and decision support and make the impact on ad performance and efficiency clear.
  3. Design how the sub-agent connects with downstream ads products (PCL, ROAS bidding) and internal tools, including APIs, contracts, and workflows that surface product‑aware alerts and ranked recommendations to identify opportunities and power performance lifts.
  4. Consolidate and structure the measurement context layer—matched and attributed conversion tables, enrichment pipelines, and existing tools—into high-quality, AI-consumable signals the agent can query and reason over.
  5. Partner with Product, Operations, Sales, and other ads product teams to translate measurement pain points into agent skills (e.g., diagnosing EQS drops, PCL readiness, partner‑specific issues) and iterate quickly on internal-first experiences before expanding to advertiser-facing use cases.

Skills

Required

  • backend or full-stack software engineering experience building large-scale distributed systems, services, and data pipelines
  • shipping GenAI- or ML-powered products end-to-end (agent or model integration, retrieval, evaluation, safety/guardrails, and online/offline metrics)
  • prior ads domain expertise, preferably in measurement, including conversion tracking, attribution, signal enrichment pipelines, and familiarity with concepts like ROAS, CPA, and campaign optimization
  • business outcomes, with a track record of tying technical work to performance, measurement quality, and operational efficiency metrics
  • product scoping, data analysis and experimentation (e.g., SQL over large datasets, experiment design, cohort analysis)
  • technical leadership across teams: scoping ambiguous problems, aligning with product and XFN partners, and driving complex initiatives from vision through launch with clear documentation and communication
  • Experience upleveling engineers in AI tooling and best practices, including setting team-wide standards, templates, or processes for building and evaluating AI-assisted workflows.

Nice to have

  • ideally in ads, measurement, or similar data-intensive domains
  • familiarity with concepts like ROAS, CPA, and campaign optimization
  • connect conversion-health interventions to performance outcomes like ROAS, CPA, and PCL validity

What the JD emphasized

  • Own the design and implementation
  • Evolve the existing GenAI Applications
  • Design how the sub-agent connects
  • Consolidate and structure the measurement context layer
  • Partner with Product, Operations, Sales
  • Lead an evolving GenAI culture
  • Required prior ads domain expertise
  • Proven track record shipping GenAI- or ML-powered products end-to-end

Other signals

  • GenAI-powered Conversion Health agent
  • reason over EQS, conversion funnels, and their impact on performance at scale
  • proactively detect issues, recommend and automate fixes
  • demonstrate measurable performance impact for advertisers