Staff Software Engineer - UI Foundation

The UI Foundation team owns the frontend infrastructure that every product engineer at Amplitude builds on — architecture, performance, developer experience, tooling, and build systems. We're the team that makes shipping fast, reliable frontend code possible at scale, and we're increasingly focused on how AI agents can autonomously handle engineering workflows that used to require constant human intervention.

The Role

We're looking for a Staff Software Engineer who thinks like an infrastructure engineer but has deep fluency in frontend systems. You'll drive architectural decisions for our frontend platform, push performance and DevX forward, and lead the build-out of our "software factory" — a set of autonomous, AI-agent-driven engineering workflows that increase engineering velocity across the org.

This is a high-leverage, low-oversight role. You'll be setting technical direction for a foundational team, not just executing a roadmap someone else wrote.

What You'll Do

• Frontend architecture: Own and evolve the architectural foundations of our frontend platform — module boundaries, micro-frontend patterns, shared component/design-system infrastructure, and state management strategy.
• Application performance: Diagnose and fix systemic performance issues (bundle size, load time, runtime performance, rendering bottlenecks); establish performance budgets and guardrails that scale across teams.
• Developer experience & tooling: Build and maintain the build systems, local dev environments, CI/CD pipelines, linting/codegen, and internal tooling that other engineers depend on daily.
• Software factory / agentic workflows: Continuously identify the biggest bottleneck in the engineering workflow and build AI-agent-driven automation to remove it, then move to the next one. For example, once code and PR generation stopped being the constraint, the bottleneck shifted to validation — and that's the next thing to automate. This is an ongoing, iterative practice of finding where engineers are spending unnecessary manual effort and systematically eliminating it with agentic tooling.
• Technical leadership: Set technical direction and standards for the frontend org; mentor senior/staff-track engineers; review architecture proposals from other teams that touch shared frontend infrastructure.
• Cross-functional partnership: Work closely with product engineering teams, design systems, and platform/infra teams to make sure foundational investments actually solve real product-team pain points.

What We're Looking For

• Real world software engineering experience, with a track record operating at Staff level or equivalent scope/impact.
• Deep expertise in modern frontend architecture (React or equivalent component frameworks) combined with strong infrastructure/systems engineering instincts — you're as comfortable in build tooling, CI/CD, and dev environments as you are in application code.
• Demonstrated experience improving frontend application performance at scale (bundle optimization, rendering performance, load-time metrics).
• Experience designing and shipping internal developer tooling or platform infrastructure that measurably improved developer velocity.
• Hands-on experience with AI-agent-driven or LLM-powered engineering workflows (agentic coding tools, autonomous pipelines, AI-assisted codegen/migration) — or a clear point of view on how to build them.
• Strong technical judgment and the ability to drive consensus on ambiguous, cross-cutting architectural decisions without heavy-handed authority.
• Excellent communication, written and spoken — this role requires influencing through documents and proposals as well as in design reviews and cross-team discussions, as much as through code.

Nice to Have

• Experience with monorepo tooling, micro-frontend architectures, or large-scale frontend platform migrations.
• Familiarity with design systems and component library architecture.
• Prior experience building "agent harness" style infrastructure: eval frameworks, agent orchestration, or tool-use frameworks for LLMs.