• You'll be responsible for leading the architecture and development of our software systems, ensuring an exceptional user experience • You'll be working in our AWS stack and helping collect, process, and display streaming sensor data in real-time to various internal and external consumers • Set and uphold the highest standards for UI/UX • Optimize and design user-facing software for high-concurrency workloads • Ensure performance under realistic loads and access patterns • Integrate frontend systems with backend services, handling backend/database work where needed • Provide thorough tests and alerting to ensure resiliency • Write well-documented code and be prepared to explain your solutions collaboratively; conduct and receive code reviews • Work directly with Engineering and Product Management, as well as fellow engineers, to arrive at the best possible solution

• Build and ship full-stack services and applications using Golang, gRPC, TypeScript, and React to support CX agent workflows and automation • Own the design and implementation of third-party platform integrations that extend the capabilities of Coinbase's internal CX tooling • Partner with team members on architecture decisions, code reviews, and engineering best practices across testing, deployment, and observability • Identify and mitigate technical risks during design and implementation to maintain platform reliability for high-traffic, production systems • Translate complex technical concepts into clear narratives for non-technical stakeholders across CX, product, and leadership teams

• Products end-to-end. You pick up a problem from product or a founder, write the spec yourself, decide what's in and out of scope, build it, ship it, and watch the metric. If the metric doesn't move, you figure out why and try again. • A direct line to product. No PM in the middle. You talk to product directly, push back when you should, and decide together what's worth building this week. • A team of agents. Claude (or whatever ships next month) handles code review, test generation, and boilerplate. You set the rules, review the reviews, and keep the bar for agent-produced code in your project above the company average. • Production health in your area. Bug count, incidents, test coverage, the long-term shape of the codebase agents are touching. You own the number, and we measure it. • Speed, measured. Lead time, throughput, agent cycle time. You'll have a personal baseline in week 2 and an improvement target by month 2. We don't count tickets closed. • Saying no. Roughly a third of what lands on your plate shouldn't get built. The right answer is often "let's run a 2-day experiment first" or "this doesn't move a metric we care about". Push back with reasoning, not instinct.

• Products end-to-end. You pick up a problem, write the spec yourself, decide what's in and out of scope, build it, ship it, and watch the metric. • A direct line to product. No PM in the middle. • A team of agents. Claude (or whatever ships next month) handles code review, test generation, and boilerplate. • Production health in your area. Bug count, incidents, test coverage, the long-term shape of the codebase agents are touching. • Speed, measured. Lead time, throughput, agent cycle time. • Saying no. Roughly a third of what lands on your plate shouldn't get built.