Confirmed on the employer's own hiring board on Nov 19, 2026. First seen by Alion on Nov 19, 2026. Google scores B on the Alion truth index.
About the job
We engineer the end-to-end lifecycle of Trust & Safety data, maintaining the production data marts that serve as our definitive single source of truth. We go beyond traditional data engineering by building intelligent DE agents and operationalizing robust, scalable pipelines. As trusted advisors on data quality, lineage, and technical feasibility, our work ensures that the data ecosystem is prepared to meet the complex, high-stakes demands of the AI era.
As a Data Engineer supporting the Responsibility, Trust & Safety Security Engineering (RT&S) Quality team, you will drive foundational, AI-native data architecture and technical solutions that safeguard Google's ecosystems and uphold policy enforcement standards across 19+ Product Areas (PAs). In this role, you will lead the evolution of the Quality Data Warehouse, moving from fragmented legacy pipelines to high-throughput, agent-ready data systems.
You will manage complex data challenges spanning multi-rater evaluation Directed Acyclic Graphs (DAGs) across manual reviews and automated autoraters, precision/recall drift, statistical sampling, and regulatory readiness (e.g., EU DSA and EU AI Act). By engineering a unified data foundation, you will automate root cause analysis (RCA), connect quality measurements to operational capacity/operating expense (OpEx), and power real-time AI/LLM evaluation guardrails. You will collaborate closely with Quality Program Managers, Data Scientists, and Applied AI teams to establish canonical quality standards across Google's operations.
Individual pay is determined by factors including job-related skills, experience, and relevant education or training.US: $156000 - $226000 (USD) + 15% bonus target + equity + benefits
Learn more about benefits at Google.
Responsibilities
- Design and scale unified, config-driven data models and SQL Pipelines for the Quality Data Warehouse.
- Map and unify multi-layered evaluation processes across disparate review environments-including multi-rater reviews, consensus audits, petitions, golden sets, and Full-Time Employee oversight-treating Human Raters and Applied AI/LLMs as unified review entities.
- Implement streaming and near-real-time data ingestion pipelines for statistical sampling, anomaly detection on precision/recall/Negative Predictive Value metrics, and model-drift detection.
- Deliver structured data layers feeding agentic tools for push-button root cause analysis and petition insights.
- Partner with cross-product area data engineering tracks to join Quality Gold data with review operations and capacity forecasting, delivering visibility into "Always-On" quality return on investment across automated and vendor workflows. Enforce Medallion data tier standards, automated data quality gates, and immutable lineage for internal compliance audits and external regulatory transparency reporting.
Qualifications
Minimum qualifications:
- Bachelor's degree in Computer Science, Mathematics, a related field, or equivalent practical experience.
- 5 years of experience with data processing software (e.g., Hadoop, Spark, Pig, Hive) and algorithms (e.g., MapReduce, Flume).
- Experience writing software in Java, Python, or SQL.
- Experience managing client-facing projects, troubleshooting technical issues, and working with Engineering and Sales Services teams
- Ability to communicate in English fluently to support client relationship management in this region.
Preferred qualifications:
- Experience designing data architectures for Trust & Safety, quality assurance, human-in-the-loop (HITL) workflows, or AI/LLM evaluation ecosystems.
- Experience transitioning legacy batch-processing pipelines to near real-time streaming architectures and AI-native, agentic workflows.
- Experience building production LLM/GenAI systems: multi-agent orchestration, agent frameworks (e.g., ADK), tool and function calling, RAG, and prompt/context engineering.
- Demonstrated track record of collaborating across cross-functional partners (Data Scientists, Software Engineers, Product Managers, and Policy/Operations teams) to drive technical standardization and deprecate legacy tech debt.
- Familiarity with statistical methodology (e.g., sampling design, confidence interval modeling, Inter-Rater Reliability/IRR metrics, anomaly detection).

