Platform Engineer

As an Agent Platform Engineer, you'll build the core infrastructure that makes our agents reliable, observable, evaluable, and scalable across customers and workflows. This is a deep technical role focused on computer-use infrastructure, agent architecture, evaluation systems, orchestration, reliability, and the internal platform that lets us deploy high-quality agents quickly. You should be energized by hard systems problems at the edge of what AI can reliably do today: browser and desktop automation, agent planning and execution, workflow orchestration, evaluation frameworks, observability, tool use, memory, failure recovery, and human-in-the-loop control. One day you might be designing an evaluation harness for agent performance. The next, you might be improving our computer-use runtime, building better debugging tools, designing new agent architecture, improving our document understanding pipeline, or creating reusable primitives for workflow execution across EMRs, portals, faxes, PDFs, and other fragmented systems. What You'll Do Design and implement infrastructure that powers computer-use agents across customer workflows — browser automation, desktop automation, tool use, state management, retry logic, human escalation, workflow executionBuild evaluation and observability systems to measure agent performance, reliability, accuracy, and failure modes; build tools for replaying runs, debugging agent decisions, and tracking regressionsExperiment with and productionize better approaches to planning, execution, memory, tool use, validation, and recoveryOwn backend systems: APIs, domain services, background jobs, workflow engines, event-driven systems, and reliability mechanisms across our TypeScript monorepoPartner closely with Forward Deployed Engineers to turn customer-specific learnings into reusable platform infrastructure What We're Looking For 4+ years of software engineering experienceStrong TypeScript experience across backend systemsStrong backend fundamentals: APIs, databases, queues/workflows, service boundaries, distributed systems, reliabilityExperience building production systems with background jobs, workflow engines, event-driven architecture, or orchestration frameworksComfort with relational databases and ORMs, ideally PostgreSQL and PrismaStrong systems design instincts and ability to create clean abstractions in ambiguous technical domainsStrong AI-assisted coding workflow — comfortable with tools like Cursor, Claude Code, Codex, or similarHigh ownership — you can take an ambiguous technical problem and drive it to a robust, maintainable solution Nice to Have Experience building AI agents, LLM-powered products, computer-use systems, or browser/desktop automationExperience with evaluation frameworks, test harnesses, or observability/debugging tools for AI systemsExperience with Temporal or similar workflow orchestrationExperience with Playwright or remote desktop infrastructureExperience with document processing, retrieval systems, or structured extraction pipelinesExperience with GCP, Docker, Pulumi, or modern cloud infrastructureExperience in regulated, high-compliance, or data-sensitive environments Visa: H1B transfers only; unable to sponsor new visas at this time.