Infrastructure Engineer, Ivo

• Who? This isn’t a “keep the lights on” role. You’ll be building the system that keeps the company running. In addition to helping us run a solid, high-performance distributed system, we’d love someone who’s as excited about LLMs as we are. You’d be deeply embedded into the engineering team and highly encouraged to push the frontiers
• You've actually lived in Kubernetes, not just deployed to it. You get cluster architecture, scheduling, networking, storage primitives, and the fun ways distributed systems fail
• IaC and CI/CD are second nature — Pulumi or Terraform, Docker, GitHub Actions, and you've done it across multi-cluster, multi-region setups, not just a single happy-path environment
• You think in failure modes. Not "will this break," but "when this breaks, what happens next" — and you design for that from day one
• You can turn "the contract says X" into "the infra does X." Compliance and legal requirements don't stay abstract with you around — you translate them into real constraints
• You're a DevOps/infra person at your core, with enough full-stack range to not get stuck when the bug crosses into backend or frontend territory
• You debug like a detective. Systematic, relentless, not afraid to go five layers deep to find the real root cause
• Ambiguity doesn't scare you. "Something's broken somewhere" is a starting point, not a blocker — you trace it across systems until you find it
• Years of experience: 2+

Desirable

• Why? Infrastructure Engineers build the foundation for Ivo’s entire platform. Customers are cagey about their contracts, so each customer gets their isolated environment with containers, database, VPC, etc. Things break. Regions go down. Cloud and LLM providers have “incidents.” Customers still expect us to hit our SLAs
• What? This list isn't a checklist, it's a direction. We don't need you to have done all of it — we need you excited to figure out the parts you haven't
• Run Kubernetes like it's your own startup within the startup — own multi-cluster, multi-region deployments across AWS/GCP/Azure, with failover and disaster recovery that actually works when it matters (not just on paper)
• Build the tools that build our infra — internal tooling for spinning up and managing clusters, so dev-to-prod is consistent and nobody's hand-crafting environments at 2am
• Figure out the right way to isolate workloads — ML vs. API traffic have very different needs; you'll design the strategy that balances cost, performance, and reliability without over-engineering it
• Make security invisible, not annoying — RBAC, workload identity, secrets management, data residency, audit trails. Enterprise customers need to trust us; product velocity shouldn't suffer for it
• Set SLOs that don't lie — define targets that are real, enforceable, and won't page you at 3am for nothing. Hold teams (including yourself) to them
• Own the pipes — GitHub Actions for CI/CD, Pulumi for infra as code. If it ships, you had a hand in it
• Keep Docker environments sane — across dev, staging, and prod, so "works on my machine" stops being a punchline
• Go wherever the bug is — APIs, workers, jobs, frontend builds and performance. You're not siloed to "infra"; you debug end-to-end
• Make everyone's day faster — faster builds, cleaner environments, less manual toil. If it's repetitive, automate it
• Own uptime like it's personal — build observability that tells you what broke, why, and how often, before a customer has to tell you first
• Lead when things go sideways — run incident response, write postmortems people actually read (and learn from)