Staff Backend Software Engineer - Activations at LiveRamp in San Francisco, CA

Job description

LiveRamp is shaping the future of responsible data collaboration between the world’s leading brands, retailers, financial services providers, and healthcare innovators. As consumers embrace new AI-driven experiences, the LiveRamp data collaboration network exponentially expands the breadth and accuracy of the data on which marketing AI capabilities operate, powering deeper customer insight and measurable performance on a global scale. LiveRamp is headquartered in San Francisco, California, with offices worldwide. Learn more at LiveRamp.com . The Activations Backend team is responsible for the bulk of big data processing that powers LiveRamp’s primary activation product, which delivers hundreds of millions in annual recurring revenue. Our systems process over one hundred thousand batch jobs per day, ranging in size from gigabytes to over 100 terabytes, and power distributions to hundreds of downstream destinations. Cumulatively, our systems process multiple exabytes per year. We provide detailed monitoring, statistics, error recovery, and resiliency to keep this core product reliable for our largest customers. At LiveRamp, big data processing is not just for back-office analytics. Our product is a big data product: we transform, deduplicate, and transport massive datasets across clouds and regions, while respecting complex rate limits and SLAs, and enabling many of the most successful companies on the planet to activate their data safely and efficiently. Your team will: Re‑architect Activations Back End to be cloud‑forward and multi‑regional, taking advantage of SingleStore and other modern data warehouses to replace legacy Spark‑heavy flows where it makes sense. Deliver high‑throughput, low‑latency activation by combining: SingleStore‑backed state and delta computation, Spark/Dataproc for the heaviest batch workloads, and Streaming infrastructure (e.g., Redpanda/Pub/Sub) for event‑driven and incremental deliveries. Build smarter orchestration and scheduling using Temporal/Cadence and queueing services to: Classify jobs (latency‑critical vs. throughput‑heavy vs. background), Route them through multi‑queue schedulers and capacity pools, respect destination‑specific rate limits and SLAs, and eliminate duplicated work through config canonicalization and caching so identical or similar jobs across customers share heavy computation. Collaborate closely with partner teams (Identity, Data Foundation, Activations Fullstack, Integrations/OPI) to deliver end‑to‑end improvements in activation reliability, speed, and observability, and continuously raise the bar on operational excellence. You will: Lead the design and evolution of a petabyte‑scale activation platform, pushing it toward a delta‑first, cache‑aware, and cost‑efficient architecture. Shape end‑to‑end technical strategy for major areas of Activations Back End (e.g., matching/delta computation, job orchestration, delivery pipelines), from design through rollout and long‑term maintenance. Architect and build big data pipelines using Apache Spark/Dataproc, SingleStore, Kubernetes/GKE, and streaming systems (e.g., Pub/Sub, Redpanda/Kafka) where appropriate. Use workflow engines such as Temporal and Cadence to orchestrate complex, long‑running workflows with robust retry, compensation, and observability, and define patterns other engineers can reuse. Design for multi‑tenant fairness and scalability, ensuring small, latency‑sensitive jobs stay fast while large backfills and bulk workflows do not starve the system, via job classification, queueing, and rate‑limit–aware scheduling. Drive performance and cost optimization for petabyte‑scale workloads: reduce duplicate processing, improve cache hit rates, tune cluster sizing and autoscaling policies, and set and track SLOs. Lead production excellence: own critical services in production, coordinate incident response and postmortems, and drive structural fixes that meaningfully reduce operational load and risk. Infuse AI into how we build