Lead Infrastructure Engineer

About Promake

Promake is a multi-tenant SaaS company serving thousands of customers, with a product that spans both a customer-facing web platform and an AI/LLM layer.

Our production stack runs on AWS a container-based platform on EKS, managed databases, search, and a global edge tier. It's a real, load-bearing environment serving live customers, not a greenfield playground.

The role

You'll own AWS end-to-end identity & access, networking, EKS, data, observability, and cost. This is a senior individual-contributor role: you set the direction for how our infrastructure is built, secured, and operated, and you do the hands-on work to get it there. You'll be the person the team trusts to say "here's the risk, here's what we fix this quarter, and here's what we consciously accept."

We're a small team, so this role carries genuine ownership from day one. There's no layer of infra above you to defer to the judgment is yours to exercise.

This is an İstanbul-based hybrid role we work together in the office 4 days a week.

What you'll do

• Own the EKS platform: control plane, networking, node lifecycle, and the reliability of everything running on it.
• Harden identity and access — IAM users, roles, OIDC/federation, service principals, and the boundaries between them.
• Own the data tier's operational posture: RDS HA/backups/encryption, S3 fleet baseline, OpenSearch.
• Run LLM inference across multiple platforms and providers, and own our inference gateway — the layer that routes and load-balances model traffic. Keep it resilient (graceful failover across providers) and performant (latency and throughput under real load) as usage grows.
• Build a focused observability story — instrument what actually needs watching, with meaningful alarms and SLOs on the critical paths, so we find out about real problems before customers do without drowning in noise.
• Keep egress and cost honest; find and kill the outliers.
• Codify it. Infrastructure-as-code (Terraform/OpenTofu) is how changes ship, with rollback and blast radius understood before apply.

What we look for

We care more about judgment than about a checklist of services you've touched. The engineers who thrive here can:

• Prioritize and defend it. Sort risk into "fix now / accept-as-risk / defer" and justify every row. A tight, well-reasoned shortlist beats an exhaustive dump.
• Reason about blast radius. For any change or gap, you think through data exposure, availability, recovery time, customer impact, and regulatory surface — not just "this is a misconfiguration."
• Discover without being told. Given a production account and a loose brief, you find the things that matter unprompted.
• Communicate so others can act. Your writeups let a reader make a decision without a follow-up meeting.

Requirements

• 4–6 years operating production AWS infrastructure at meaningful scale.
• Deep, hands-on EKS/Kubernetes — control plane, networking, node management, not just kubectl
• apply .
• Strong IAM and network security fundamentals.
• Fluent in Terraform or OpenTofu, with real opinions about module structure, state, and safe rollout.
• Comfortable owning RDS, S3, and observability tooling in production.
• Experience operating high-throughput proxy/gateway or API-routing infrastructure where failover and latency matter (LLM-serving experience is a strong plus).
• Can write clearly for a technical audience.

Nice to have

• Multi-tenant SaaS experience.
• Hands-on with LLM inference across multiple providers/platforms. CloudFront/edge and multi-region experience.
• Cost-optimization track record.



Our hiring process

We run a ~4-hour take-home: a breadth audit of a real (read-only) slice of our production AWS account, plus one deeper artifact of your choice (an IaC remediation module, or an on-call runbook + monitoring proposal). We share the rubric and its weights up front — there's no hidden scoring trick. We're evaluating judgment and prioritization, not coverage.