Senior Software Engineer II - Action Infrastructure - Action Runtime

DigitalOcean•Seattle, WA

6d•Hybrid

About The Position

DigitalOcean is seeking a Senior Software Engineer II for the Action Runtime team. This team is responsible for building the execution layer for AI agents, including tool execution environments and orchestration systems that enable agents to run code and take actions safely, quickly, and at scale. The role involves defining standards for agent execution, such as isolation guarantees, cold-start latency, and throughput, and then designing, building, launching, measuring, and operating these systems end-to-end. This is a senior individual contributor position for a self-starter who can independently translate ambiguous customer and business needs into product planning, design, technical design, production implementation, launch, and long-term maintenance. The ideal candidate is comfortable working full-stack, collaborating across product, design, and engineering, and making pragmatic technical choices within fast-moving agent tooling ecosystems.

Requirements

Senior-level experience designing, building, shipping, and operating production software systems with meaningful reliability, scale, security, or customer impact.
Production experience with Go, TypeScript, and Python, or deep expertise in one or two with the ability to work effectively across all three.
Hands-on experience building with or deeply using agent and AI coding tools such as Claude Code, Codex, Cursor, or similar systems.
Practical familiarity with agent tool design, including MCP, tool contracts/schemas, tool selection, model selection, agent evaluations, observability, and failure analysis.
Full-stack ownership mindset: you can build backend services and APIs, shape developer or operator workflows, and make the product experience work end to end.
Strong production engineering instincts for cloud systems: SLOs, incident response, monitoring, tracing, capacity, deploy safety, auth, secrets, abuse prevention, and supportability.
Ability to turn ambiguous requirements into clear RFCs, architecture docs, implementation plans, milestones, and tradeoff decisions.
Experience influencing cross-functional projects through clear written communication, technical judgment, and close partnership with PMs, designers, managers, and peer engineering teams.
A track record of mentoring or coaching other engineers through technical decision-making, code quality, architecture, and written/verbal communication.

Nice To Haves

Experience building tool ecosystems, integration platforms, developer platforms, workflow automation, or cloud control-plane systems.
Experience evaluating LLM or agent behavior in production-like environments, including regression suites, offline/online evals, or quality scoring.
Familiarity with third-party API reliability patterns such as idempotency, pagination, retries, backoff, webhook verification, OAuth, quota handling, and vendor failure modes.
Contributions to agent tooling standards, open-source tools, developer communities, or technical talks.

Responsibilities

Independently own and ship product features within Action Infrastructure from planning through design, implementation, launch, operations, and maintenance with high autonomy.
Build production-grade third-party agent tools and supporting infrastructure, including MCP servers, tool schemas, authentication and secrets handling, rate limiting, observability, and safe execution boundaries.
Create full-stack experiences for builders and operators to discover, configure, test, debug, and monitor agent tools.
Collaborate closely with product, design, and UX partners to make agent-tool workflows understandable, trustworthy, and efficient for builders and operators.
Define and improve quality bars for tools using evaluations, golden tasks, telemetry, error analysis, model and tool selection experiments, and production feedback loops.
Translate business and customer requirements into clear, minimally complex technical designs that multiple teams can understand and execute against.
Partner with product managers, designers, infrastructure teams, security, support, and other engineering teams to deliver reliable customer-facing capabilities.
Guide technical decisions for a group of engineers, review code and architecture, and raise the team's bar for reliability, maintainability, and operational excellence.
Operate what you build: participate in on-call practices, debug multi-system production issues, lead incident diagnosis when needed, and turn operational learnings into durable improvements.
Track emerging agent tooling standards and practices, especially around MCP, evals, tool-use reliability, agent UX, and AI coding/agent environments, and apply them when they create customer or business value.