As a Data Scientist — Agent Evaluations & Quality, you will own the measurement system that determines whether the assistant is genuinely improving in ambiguous, real-world environments.
You'll partner directly with AI Agent Capabilities engineers to generate the evidence that shapes product decisions, model choices, and release quality.
This is a high-ownership, deeply technical role at the intersection of applied data science, LLM evaluation, and product quality — ideal for someone who thrives on turning hard, open-ended quality questions into rigorous, actionable answers.
What You'll DoArchitect and maintain automated evaluation pipelines that measure agent quality across product surfaces.
Translate agent capabilities into explicit pass, partial-pass, and failure criteria for complex multi-step tasks.
Build representative gold datasets and regression suites covering real workflows, edge cases, and adversarial scenarios.
Define meaningful metrics — task success, tool-selection accuracy, instruction adherence, factual consistency, latency, cost, and reliability.
Design deterministic and model-based graders, calibrate LLM-as-a-judge systems, and track grader agreement.
Compare models, prompts, and implementations using rigorous offline experiments and production evidence.
Analyze traces and production outcomes to identify root causes and build a practical failure taxonomy.
Turn production failures into regression cases and continuously close gaps in evaluation coverage.
Build dashboards and release-quality signals that make results actionable for engineering, product, and leadership.
Recommend improvements to capability engineers and verify that fixes raise quality without unacceptable regressions.
What We're Looking ForRequired4+ years in Applied Data Science or Machine Learning roles, with a track record of building and delivering evaluat.