Want to know if this job is worth applying to?
Mountain View, California, United States
Hybrid
Help develop the technical roadmap and architecture for the evaluation platform, from offline benchmarking to online/production monitoring of agentic and LLM-based systems.
Design eval methodologies appropriate to different stages of the pipeline: golden/regression test sets, human-in-the-loop review workflows, LLM-as-judge approaches, and automated metrics for task success, safety, and hallucinations.
Build the data infrastructure evaluation depends on: annotation and labeling pipelines, dataset versioning, data quality checks, and tooling that lets researchers and product teams run and interpret experiments without needing platform team help.
Partner closely with Research, Product, and Platform teams to productize experiments into robust AI solutions
Represent the eval platform to stakeholders outside the immediate team- set expectations on what "good" looks like for a model/agent release, and report on platform health and coverage.
Stay current with advancements in ML, NLP, voice, and LLM systems, and contribute actively to technical discussions across teams.
Mentor and support other engineers through design reviews, feedback, and knowledge sharing.
Deep, hands-on experience building and operating evaluation systems for modern ML/LLM/agentic systems- not just consuming existing eval tools.
Demonstrated experience leading the technical direction of a project or small team: setting architecture, driving design reviews, and being accountable for a system's long-term health (not just shipping features).
Strong architectural skills, with proven experience designing complex, data-intensive software systems and production experience with Python, AWS, Kubernetes, and/or Docker.
Experience designing data pipelines for ML evaluation- labeling/annotation workflows, dataset versioning and quality control, and reproducible benchmarking.
A Bachelor’s Degree in CS or other related fields
Demonstrated technical mentorship of junior and mid-level engineers, driving adoption of best practices and architectural alignment for scalability and extensibility.
Desire to learn, teach, and collaborate closely with cross-functional peers.
Experience building and evaluating agentic systems at scale.
Experience with voice/audio quality evaluations.
Production experience with LLM-centric services (e.g., inference, orchestration, evaluation, monitoring)
Familiarity with large-scale ML experimentation, benchmarking, or simulation frameworks.
Experience with conversational/customer-support AI domains (e.g., containment rate, conversation quality, goal completion).
Knowledge of techniques for optimizing model architectures for faster inference.
Experience with AWS, CI/CD, Kafka, Athena
