Architect canonical, evolution-safe telemetry agreements and schemas building on integration event protos, trajectory schemas, and AI-assisted instrumentation across developer and DevAI surfaces so new agentic tools emit structured, high-signal telemetry by default. Design and build low-latency ingestion, deterministic sampling, Personally Identifiable Information (PII) redaction, and LLM-powered trajectory classification pipelines.
Build core attribution engines linking multi-turn agent trajectories, tool calls, skill invocations, token spend, and downstream artifacts including Piper CLs, Critique reviews, Tracker issues, and test runs to power Levels of Autonomy and end-to-end Critical User Journeys. Establish engineering standards across the data and serving stack to eliminate pipeline fragility and enforce data freshness and completeness SLOs.
Partner with Engineering Directors, Principal Engineers, Data Scientists, and sibling DevAI/CodeAI platform leads to unblock cross-organization instrumentation bottlenecks, review critical CLs and RFCs, and mentor engineers through example.
Job description provided by Google.
View the original posting
