Principal Product Software Engineer, Equinix

• 12+ years of software engineering with a record of technical leadership across teams, without needing a management title to lead
• Deep experience designing and operating distributed systems: APIs, event-driven architecture, cloud-native platforms
• Hands-on experience building LLM-powered applications or agents in production, including evals, guardrails, and failure analysis
• Fluency in Python or TypeScript and the craft to review anyone's code credibly
• A record of setting technical standards others adopted, and of mentoring senior engineers
• Clear, direct technical communication, from whiteboard to executive summary

Desirable

• Equinix Incubation builds the agentic systems that run the software delivery lifecycle end to end: specialized AI agents, with humans directing the work and owning every gate
• Principal Engineer is the senior-most hands-on role on that mission
• We are hiring for two assignments with one bar
• The platform lead owns the orchestration backbone: the delivery orchestrator, agent-to-agent protocols, eval harness, and the platform's reliability, security, and cost at scale
• The agent tech lead owns the technical direction for the agent fleet across intake, portfolio, requirements, and operations: how each agent is designed, evaluated, hardened, and trusted
• Both write code
• Both set the standard everyone else builds to
• Design the architectures that make an agent fleet dependable: orchestration, A2A handoffs, MCP tool use, event-driven state, and enterprise integration
• Take the problems no one else can unblock (agent reliability, eval design, latency and cost at scale) and return them solved, documented, and reusable
• Write and review code where it matters most; your fingerprints should be on the foundations
• Define eval standards, guardrails, human-in-the-loop thresholds, and release criteria for agents, and make them stick across pods
• Build the eval harness and observability that turn agent quality from opinion into measurement
• Own technical sign-off on agent and platform designs, with the judgment to know when good enough is good enough
• Mentor engineers into agent engineering: prompt and context patterns, tool design, failure analysis, and taste
• Establish reusable patterns, reference implementations, and standards that make the second agent cheaper than the first
• Partner with the Director and Product Manager to turn business intent into build-ready technical direction
• Treat security, privacy, auditability, and access control as design inputs, working with InfoSec and AI governance from day one
• Engineer for operations: monitoring, rollback, capacity, and cost, so trust in the platform compounds instead of eroding
• Dogfood relentlessly: use agents to build the agent platform, and feed what you learn back in
• What Success Looks Like
• The platform's reliability, security, and cost hold as agent volume grows, because the foundations you designed anticipated scale
• Agents ship faster each quarter, on patterns and standards you established, without the quality bar moving
• Engineers cite you as the reason they leveled up; the hardest problems find you and leave solved
• Design reviews are calmer because your eval and release standards answer most questions before they are asked
• Agent Platform & Orchestration Lead: owns the orchestrator, protocols, eval harness, and platform hardening for the delivery engine
• Agent Tech Lead (Discover, Operate, Deliver): owns technical direction, design quality, and trust for the agent fleet built by the pods
• Architecture at Scale
• Distributed systems; orchestration design; enterprise integration; cost and reliability engineering
• Agent Engineering
• LLM application architecture; eval design; guardrails and HITL gates; failure analysis
• Technical Leadership
• Standards-setting; design sign-off; mentoring; influence without authority
• Craft
• Production-grade code; security by design; observability; pragmatic judgment