Principal Software Engineer, Metrics Platform

Role Overview

Roblox’s data infrastructure powers analytics, machine learning, and product decisions across a platform serving more than 200 million daily active users. As a Principal Software Engineer in Data Infra, you will set the technical direction for a metrics platform that helps teams define trusted metrics, discover and govern them, and use them consistently across analytics and data-processing systems.

You will lead architecture and execution for a platform that connects metric definitions to their materializations and consumers, including Spark, Flink, OLAP systems, and feature platforms. This role requires both deep technical judgment and organizational influence: you will align teams around durable interfaces and semantics, guide major cross-team initiatives, and help Roblox evolve how it builds and uses metrics at scale.

Responsibilities

• Set the technical vision: Define the long-term architecture and evolution of the metrics platform, including metric registration, discovery, versioning, ownership, dependencies, and governance.
• Establish shared semantics: Design the contracts and standards that let analytics and compute systems consume metric definitions consistently, without coupling those definitions to one execution engine.
• Connect definitions to data: Guide the architecture for turning logical metric definitions into physical materializations and serving locations across Spark, Flink, OLAP, and feature platforms.
• Lead cross-team initiatives: Build alignment across Data Infra, data engineering, analytics, and ML teams; resolve architectural boundaries and dependencies; and drive adoption of shared platform capabilities.
• Raise reliability and efficiency: Set expectations for scalability, observability, resilience, data correctness, and operational ownership across the metric lifecycle.
• Develop technical leaders: Mentor engineers, improve technical decision-making across teams, and create leverage through reusable designs, APIs, and engineering practices.
• Shape platform interaction: Explore self-service and agent-assisted workflows for metric discovery and use, with appropriate validation and governance.
• Agentic Interface: Embed AI/ML capabilities within our data platform, leveraging LLMs for data discovery and generation, building autonomous, self-serve mechanisms for platform interaction layer.

Qualifications

Required

• B.S. equivalent in CS or sufficient experience.
• 10+ years of experience building, designing, testing and maintaining production-grade, large-scale distributed systems.
• Extensive experience designing and delivering production-grade distributed systems or data platforms.
• A record of setting technical direction for systems that span multiple teams or product areas.
• Deep experience in one or more areas such as data platforms, metadata and catalog systems, semantic layers, analytical systems, or large-scale data processing.
• Strong engineering skills in Java, Go, Scala, or Python, and experience designing reliable services and APIs.
• Strong understanding of SQL and familiarity with systems such as Spark, Flink, Trino, Druid, or comparable technologies.
• Demonstrated ability to turn ambiguous, cross-team problems into an architecture and execution plan that teams can adopt.
• Excellent technical communication and influence: ability to build alignment, explain tradeoffs, and guide decisions without relying on direct authority.
• Cross-Organizational Technical Leadership: A proven track record of influencing technical direction across a large engineering organization, leading consensus on complex architectural initiatives, and championing successful multi-quarter projects.
• Extreme Ownership: Radical accountability, actively defining technical requirements, unblocking dependencies, rallying engineers across pods, and steering complex projects to production stability.

Preferred

• Experience building metric or semantic platforms, data catalogs, lineage systems, or data governance capabilities.
• Experience making metric definitions usable across multiple query or compute engines.
• Contributions to open-source data infrastructure projects.
• Experience operating data infrastructure at consumer-internet scale.

Work Arrangement

Roles that are based in an office are onsite Tuesday, Wednesday, and Thursday, with optional presence on Monday and Friday.