Build the Agents That Build Our Software 

Most engineering roles now come with AI tools. This one comes with a mission. Equinix Incubation builds the agentic systems that run the software delivery lifecycle end to end: from intake and business case through design, code, test, release, and value tracking, with humans directing the work and owning every gate. 

We are hiring Staff and Senior Staff Engineers to build those agents and the platform they run on. You will not just use AI to code faster. You will design, ship, evaluate, and harden production agents that colleagues across a global organization trust with real delivery work. Your agents will write software; your engineering decides what ships. 

What You'll Do 

Build Production AI Agents 

• Design, build, and ship LLM-powered agents that execute real lifecycle work: intake triage, estimation, requirements, technical design, coding, testing, release, and operations. 

• Engineer the scaffolding that makes agents dependable: tool use via MCP, agent-to-agent handoffs (A2A), event-driven orchestration, and deep Jira and enterprise system integration. 

• Build on Equinix's enterprise AI platform: AI gateway, orchestration, audit, and access control, with security and privacy by design. 

Make Agents Trustworthy: Evals, Guardrails, Gates 

• Design and automate eval suites that measure agent output quality on every change, and make passing evals the release gate for agents. 

• Define guardrails, human-in-the-loop approval points, review thresholds, and escalation paths, so agent autonomy is earned, not assumed. 

• Instrument agent behavior end to end (quality, latency, cost, adoption), find failure patterns, and tune prompts, context, and configurations until the numbers move. 

Engineer Context and Knowledge 

• Build the knowledge layers agents depend on: retrieval over process libraries, decision histories, code, and delivery data. 

• Establish reusable prompt patterns, context standards, and agent configurations that other teams adopt. 

• Own agents through their full lifecycle: instructions, context freshness, performance monitoring, feedback, and retirement. 

Ship the Platform and Raise the Bar 

• Contribute to the orchestrator, persona consoles, and dashboards that keep humans in command of agent-led delivery. 

• Dogfood relentlessly: use agents to build agent systems, and feed what you learn back into the platform. 

• Bring strong engineering craft. The fundamentals still decide whether this works: architecture, code quality, testing, CI/CD, and cloud-native design. 

What Success Looks Like 

• Agents you built are doing live delivery work, with measurable cycle-time and quality gains, and humans confidently in control. 

• Your eval suites are the reason people trust agent output; “passes evals” means something because you made it mean something. 

• Your context patterns, guardrails, and agent standards are reused by teams you have never met. 

• You can explain to an executive, in plain language, what an agent did, why, and how you know. 

• The platform gets simpler, faster, and cheaper as it scales, because you treat agent cost and reliability as engineering problems. 

Level Expectations 

• Staff: You deliver complete agents and platform components within established patterns, own their evals and quality end to end, and are the dependable engine of your pod. 

• Senior Staff: You set the patterns. You take the hardest, most ambiguous problems (orchestration, eval design, agent reliability at scale), define the standards others follow, and multiply the team. 

Required Qualifications 

• 6+ years (Staff) or 9+ years (Senior Staff) of professional software engineering experience, with a record of shipping and operating production systems. 

• Hands-on experience building LLM-powered applications or agents: prompt and context engineering, tool calling, retrieval, or multi-agent workflows. 

• Experience designing evaluations for AI systems, or strong test-engineering instincts you are eager to apply to non-deterministic software. 

• Strong proficiency in Python or TypeScript, plus solid API, microservices, and event-driven architecture skills. 

• Fluency with modern engineering practice: Git, automated testing, CI/CD, observability, and cloud platforms. 

• Sound judgment about when to trust automation and when to demand human review, and the communication skills to explain that reasoning. 

Preferred Qualifications 

• Experience with agent frameworks and protocols such as MCP, A2A, Anthropic or OpenAI APIs, Bedrock, Vertex, or LangGraph. 

• Experience building developer platforms, orchestration systems, or SDLC tooling, including Jira, GitHub, or ServiceNow integration. 

• Knowledge-engineering experience: retrieval systems, embeddings, or enterprise knowledge graphs. 

• Experience taking AI features through security, privacy, and responsible AI review in an enterprise. 

• Evidence of craft: open-source contributions, technical writing, or internal platforms with devoted users. 

Core Competencies 

Agent Engineering 

• LLM application architecture; prompt and context engineering; tool use and orchestration; multi-agent design. 

Evals and Trust 

• Eval design and automation; guardrails and human-in-the-loop gates; AI observability; responsible AI governance. 

Platform Craft 

• API and event-driven design; CI/CD and automation; cloud-native engineering; enterprise integration. 

Judgment and Impact 

• Systems thinking; pragmatic risk-taking; mentoring and standards-setting; clear communication. 

Why This Role 

Incubation is a durable capability, not a project team: the team persists, and the product rotates. Agentic delivery is product one; the next incubation bets follow. You will help define how AI-first engineering works at Equinix, with the autonomy of a startup and the reach of a global platform company. Few roles let you change how an entire organization builds software. This one exists to do exactly that.