Staff Software Engineer, Backend - Platform (Core AI Automation)

As a Staff Software Engineer on the Core Automation team within the Platform group, you'll architect and build the Agentic AI systems that are transforming how Coinbase operates. This team is reimagining customer support and compliance processes for a fully AI-driven world, designing intelligent agents, orchestration frameworks, and measurement systems that deliver delightful customer experiences at scale. You'll own the technical direction for production AI systems, working across cross-functional teams to bring this vision to reality while building primitives that scale automation across the company.

What you'll do:

• Architect and build Agentic AI systems that power Coinbase's customer support and compliance automation, from intelligent agents through orchestration and guardrails
• Design foundational APIs and measurement frameworks that ensure AI agents are grounded, relevant, and reliably deliver customer delight with minimal hallucination
• Lead technical direction for distributed systems underpinning AI automation, defining architecture patterns and strategic roadmaps in partnership with engineering leadership
• Build reusable primitives and orchestration solutions that enable AI-powered automation to scale across multiple domains beyond the initial customer support and compliance focus
• Mentor engineers on AI system design techniques, coding standards, and production-readiness practices for LLM-based applications

Required skills and experience:

• 8+ years in software engineering with demonstrated experience architecting distributed systems and shipping production AI/ML applications
• Proven track record getting LLM-based applications to production, including measuring and improving grounding, reducing hallucinations, and building guardrails for reliability
• Deep expertise designing APIs, orchestration layers, and evaluation frameworks for agentic AI systems at scale
• Experience in customer support automation, compliance, or financial crimes domains with an understanding of the unique reliability and correctness requirements these areas demand
• Utilizes generative AI responsibly, maintaining human oversight to deliver business-ready outputs and drive measurable improvements in workflow efficiency, cost, and quality.