As a Staff Software Engineer on the AI Platform team within the Platform group, you'll build and operate the LLM and agent infrastructure that every team at Coinbase depends on. This team owns the company's single path to large language models and the full agent lifecycle: build, deploy, run, observe, and improve. You'll lead multi-quarter technical initiatives across the platform, from gateway and runtime systems to knowledge bases and applied AI agents, directly shaping how Coinbase scales AI across the organization.
- Own the architecture and delivery of core platform systems including the LLM Gateway (60+ models, auth, PII redaction, fallbacks, cost optimization), AI Hub, and agent runtime with microVM sandboxes and governed MCP gateway
- Drive the design and implementation of Knowledge Base infrastructure, connecting data sources to auto-provisioned vector and markdown stores queryable by any agent
- Lead AI FinOps capabilities including spend attribution, governance, and cost optimization across all AI workloads company-wide
- Partner across engineering, security, legal, finance, product, and external partners at frontier labs and major cloud providers to ship high-impact platform capabilities
- Build evaluation and observability tooling including LLM-as-judge harnesses, full tracing, and feedback loops that let subject matter experts refine production agents
- Shape applied AI strategy by embedding with teams to build production agents, applying fine-tuning and traditional ML where it outperforms prompting
- 8+ years of software engineering experience with at least 3 years building ML/AI infrastructure, LLM systems, or distributed platform services at scale
- Demonstrated ability to lead multi-quarter, Staff-level initiatives end-to-end, from technical design through production operation, with measurable platform-wide impact
- Deep expertise in LLM orchestration patterns, agent frameworks, vector databases, and model serving infrastructure across multiple providers
- Proven track record of cross-functional partnership with security, legal, and product teams to ship governed, production-grade AI systems
- Fluency in Python and at least one systems language (Go, Rust, or C++), with production experience building APIs, microservices, and runtime environments
- NodeSmith: AI-Driven Automation for Blockchain Node Upgrades
- CB-GPT - The opportunity and the vision for GenAI at Coinbase
- Lessons from launching Enterprise-grade GenAI solutions at Coinbase
- AWS re:Invent 2025 - Scaling support, compliance, & productivity with conversational AI at Coinbase