Lead Data Engineer, Kargo

• You've designed and owned large-scale, interdependent data systems, including at least one you built from scratch, and you turn ambiguity into a sequenced roadmap with Product and Data Partnerships
• You've led engineers, setting direction, reviewing work, developing people, while staying hands-on
• You have mastery of Python, Airflow and Spark, and write transformations that are idiomatic, testable and tuned for cost and performance; you write SQL for Snowflake with the same discipline
• You're at home in AWS and Kubernetes, can read infrastructure logs to diagnose failures and slowness, and have worked with third-party APIs inside ingestion pipelines
• You're fluent with AI tooling in your own work, and you think about what makes a codebase legible to it

Desirable

• Identity is central to how AdTech works today: advertisers want cross-surface reach, user-level measurement, and lower-funnel attribution. Their data reaches us two ways, and each has a clear next step: direct onboarding currently runs on a partner's identity spine and we want to build a new, multi-source graph of our own; DMP feeds are proven and now need to scale as we enter in-app inventory
• In this role you'll own the graph, a standardized ingestion path that stands up each new feed faster and cheaper than the last, and the audience state and reporting behind self-serve discovery. You'll also set the technical bar for the engineers building it with you
• Build a new identity graph . Take stock of what we have today, set its direction, and sequence the rollout: identifier sync, translation, clustering (with Data Science), opt-out handling
• Standardize partner and client onboarding across web, CTV and mobile identifiers, including cleanroom onboarding, so each new feed costs less to stand up than the last
• Ready the identity audience data layer for self-serve : creation, activation, state, and the reporting clients will discover audiences through
• Own and raise the bar on the domain's observability and alert response . Inventory today's signals, monitors and alerts, centralize them, and bring each to standard: a freshness and quality commitment, context for AI-assisted triage, and a runbook
• Lead and grow the domain's data engineers . Define the standards for testability, cost efficiency and the patterns worth repeating, then raise the team to them through your own code, reviews, and knowledge-sharing, recording decisions in ADRs