Generalize horizontal tools, including a plug-in Knowledge Store and a Self-Reflection module, into reusable services that any agent can adopt to enable automated learning and improvement.
Apply validated design principles to transition deterministic logic from monolithic prompts into efficient, code-based workflows that maximize reasoning value and ensure precise context control.
Conduct deep architectural reviews and automated trace scanning for high-impact agents to eliminate structural inefficiencies and redundant tool calls in multi-turn reasoning chains.
Extend automated prompt optimization frameworks to support multi-step workflows, enabling systematic improvement of both efficiency and quality across the agent ecosystem.
Implement code-level safety enforcements and structured tool return patterns within the agent platform to ensure reliability and correctness on deterministic subtasks.

Senior Software Engineer, Generative AI, Core ML

Job Description