Senior Software Engineer, Agent Oversight

About the role

As a Software Engineer on Agent Oversight, you will build the platform infrastructure that lets our production agents be observed, evaluated, and improved at scale. This includes building observability tooling, evaluation harnesses, and the pipelines that connect them to improvement loops. Whether building foundational infrastructure or partnering closely with ML engineers on production workflows, you will own your systems end-to-end while maintaining rigorous technical standards.

• You will:
• Design and build core platform capabilities for deploying, monitoring, and evaluating agentic applications in production
• Build reliable APIs and data pipelines that capture agent telemetry, evaluation signals, and performance metrics at scale
• Work alongside ML engineers where platform work intersects with evaluation or improvement systems — bringing enough ML fluency to reason about model behavior, evaluation quality, and improvement loops while owning the software systems that make those workflows reliable
• Own the reliability, scalability, and observability of platform components serving multiple concurrent enterprise and government customers
• Work cross-functionally with product, forward deployed engineering, and customers to translate real-world deployment requirements into platform features
• Build features end-to-end: system design, implementation, debugging, and testing
• Participate in high-velocity experimentation to validate platform capabilities against real customer usage

Requirements

• 4+ years of professional software engineering experience, with strong fundamentals in backend/distributed systems, APIs, and data pipeline design
• Hands-on experience building production software for ML/LLM-powered products or platforms, such as evaluation systems, observability/monitoring, experimentation infrastructure, agent runtimes, model-serving-adjacent services, or telemetry/data pipelines
• Working knowledge of how LLM or ML systems behave in production: evaluation signals, failure modes, prompt/tool-calling workflows, experiment results, data quality issues, and the tradeoffs between offline evals and live customer behavior
• Experience partnering closely with ML engineers or applied researchers to turn prototypes, eval loops, or model-improvement workflows into reliable platform capabilities, without needing to own model training, modeling strategy, or research direction
• Experience building infrastructure or platforms that other engineering teams build on top of (internal platform, developer tools, or similar)
• Track record of taking ownership of features or components end-to-end — from design through production — within a larger platform or system
• Comfortable operating in an ambiguous, fast-changing domain where tooling and best practices are still being defined
• Strong problem-solving skills and the ability to work independently or as part of a tight-knit, cross-functional team
• Excited to work directly with ML engineers and customer-facing teams, including challenging assumptions in designs and metrics when platform behavior, model behavior, and customer needs intersect
• Gives direct, substantive feedback on designs and code, and takes it the same way — and mentors others as they grow