Staff Software Engineer (AI Agent Evaluations)

• This Staff Engineer role sits at the intersection of engineering, applied AI, testing and developer experience. You will define and lead the discipline of testing AI agents, evaluating LLM behavior, and ensuring the reliability of agentic systems operating in production
• It requires deep engineering rigor, original thinking about what “correctness” means for non-deterministic systems, and the ability to build eval infrastructure and developer tooling that the entire engineering org depends on
• Expert in building and maintaining Retrieval-Augmented Generation (RAG) pipelines, with a deep focus on strategic data chunking and data quality enforcement
• Experience in establishing pre-retrieval data quality gates to optimize vector search accuracy, minimize retrieval-induced noise, and significantly reduce LLM hallucination rates in production-deployed agent systems
• You will establish quality standards for how ID.me ships AI-powered features safely, mentor engineers across teams on AI testing best practices, and partner directly with product and platform teams to embed quality into every stage of agent development
• Define AI Quality Standards: Own the framework for how ID.me evaluates, validates, and monitors AI agents — from prompt-based features to fully autonomous multi-step workflows
• Build Eval Infrastructure: Design and maintain evaluation pipelines for LLM outputs, agent behavior, tool use, and multi-turn interactions across development, staging, and production environments
• Production Observability for Agents: Instrument agentic systems for behavioral drift, regression, and failure modes that traditional metrics miss — latency, correctness, hallucination rate, tool misuse, and policy adherence
• Agentic Test Strategy: Lead the design of test suites that handle non-determinism — red-teaming agents, golden dataset construction, LLM-as-judge pipelines, and property-based testing for AI outputs
• Champion Developer Experience: Build the internal tooling, feedback loops, and testing workflows that make it fast and safe for engineers to develop and ship AI features with confidence
• Reduce friction in the agent development inner loop — local testing, fast eval runs, and clear signal on regressions
• Drive AI-First Engineering Culture: Raise the quality bar across the engineering org by establishing patterns, tooling, and education for how teams write, test, and deploy AI features responsibly
• Cross-Team Collaboration: Partner with Security, Platform, Product, and AI/ML teams to embed quality gates into agent development workflows
• Mentorship: Guide senior and mid-level engineers through evaluation design, observability strategy, and testing approaches specific to AI systems

Benefits

• Company-Sponsored Medical, Dental, and Vision Plans for You and Your Family
• Paid Parental Leave
• Generous Paid Time Off
• 401K With Employer Matching
• Company-Sponsored Life Insurance
• Employee Equity Incentive Plan
• Remote Opportunities
• Multiple Lounge Areas and Game Room
• Numerous Mental Health Resources- Demonstrated experience evaluating or testing LLM-powered features or autonomous agents in production
• Experience designing test infrastructure, CI/CD quality gates, or evaluation pipelines at scale
• Strong written and verbal communication across engineering, product, and leadership
• 8+ years building and operating production software systems
• Strong backend engineering fundamentals in Python, Java, Go, or equivalent
• Experience building eval frameworks for LLM agents (e.g., correctness graders, LLM-as-judge, human-in-the-loop evals, benchmark dataset curation)
• Bachelor’s degree in Computer Science, Engineering, or equivalent experience
• Red-teaming or adversarial testing experience for AI models or agents
• Experience improving developer experience — building internal tooling, reducing toil, or accelerating engineering workflows
• Familiarity with agentic frameworks (Claude API / Anthropic SDK, BrainTrust, LangChain, LangGraph, CrewAI, or similar)
• Production monitoring experience for AI systems: behavioral drift detection, output sampling, shadow scoring
• Proven ability to lead cross-team technical initiatives and influence engineering standards
• Proficiency with AI-assisted development tools (Claude Code, Cursor, or equivalent) — you build with AI every day
• Background in identity verification, fraud detection, or regulated industries
• Familiarity with Anthropic’s model evaluation methodology or similar published eval research
• Experience with observability tooling (Datadog, OpenTelemetry) applied to AI workloads
• Track record of building developer tooling or platforms that other teams adopt widely