• Define and track success metrics for AI-assisted Temporal development (benchmark scores, activation rates, time-to-production) • Build and iterate on a "SWE-Bench for Temporal" - a benchmark evaluating LLM performance on real Temporal development tasks • Create, test, and validate context files (agents.md, Cursor rules, Claude skills) that improve how coding assistants write Temporal code • Research and prioritize which LLM performance gaps to address based on user impact • Drive fast hypothesis-test-iterate cycles to continuously improve AI coding assistant performance • Partner with DevRel to ensure documentation, samples, and content are optimized for LLM consumption • Collaborate with AI Engineering on benchmark infrastructure and evaluation pipelines • Work with external AI assistant vendors to improve Temporal support • Instrument developer journeys to understand where LLM-assisted users struggle and close feedback loops