Staff ML Software Engineer (L6)

What you'll do

• Design and operate subsystems for observability, evaluation, and tooling for next-generation ML architectures.
• Build observability primitives to provide visibility into model behavior, training pipeline health, and serving latency.
• Identify and drive cost optimization across AI training and serving infrastructure through automated frameworks.
• Architect reliability improvements to reduce toil and improve on-call ergonomics across the AI/ML stack.
• Contribute to the target architecture and migration path for the modernized AIMS AI/ML stack.
• Evaluate emerging infrastructure patterns and model paradigms to develop a forward-looking platform roadmap.
• Develop subsystems for advanced agentic architectures, including memory, trace, eval, and replay pipelines.

What we're looking for

• Significant experience designing, building, and operating production AI/ML systems at scale, including training pipelines and model serving.
• Hands-on experience building subsystems for advanced agentic architectures, such as memory, trace, eval, and replay pipelines.
• Strong software engineering fundamentals with deep Python expertise and working proficiency in Scala or Java.
• Proven track record of improving AI/ML system reliability, reducing infrastructure costs, and improving operational scalability.
• Experience building observability and monitoring systems for AI/ML workloads across training, serving, and data pipelines.
• Strong distributed systems background, including batch processing at scale and real-time serving infrastructure.
• Ability to collaborate with partner teams to drive technical programs and build consensus without formal authority.
• Familiarity with LLM evaluation, trace, or replay tooling (preferred).