Senior/Staff Platform Engineer (m/f/x)

Meyandy LLC; Berlin; Vor Ort; EUR 110.000 - 170.000; Vollzeit

Zusammenfassung

Cortea AI in Berlin is seeking a Senior/Staff Platform Engineer to own the platform architecture and core stack, choosing technologies and guiding implementation from design to production.

You will work across product engineers, DevOps, and SRE to build reliable, observable systems, with emphasis on security, compliance, and scalable AI tooling.

Aufgaben

• Own the architecture and core stack for the platform and define standards for security and observability.
• Define SLIs/SLOs for critical workloads and manage the Kubernetes cluster and cloud footprint.
• Build and enforce IAM, secret management, and compliance controls (SOC 2, ISO 27001).
• Develop AI dev tooling and ensure product engineers have solid guidelines and automation to stay on the paved road.

Jobbeschreibung

Your Role

We ship fast, and we intend to keep shipping fast. The consequence is that our system is outgrowing the amount of deliberate design that went into it. We run AI agents over customer documents in a domain where being wrong is expensive, and a lot of the load-bearing decisions about how those fit together are still implicit. You will be making those decisions explicit and then making them hold. That means codifying golden paths, building the shared primitives and APIs that product engineers work on top of, owning the architecture of the system as a whole, and establishing what our reliability actually needs to be before an incident establishes it for us. Platform at Cortea sits between software engineering, DevOps and SRE. Its customers are other engineers. This is not a DevOps role under a different name: you will spend more time in application code than in YAML, and the reason we want infrastructure experience is that we don't believe you can design a system well without being able to run it. The bar we are hiring against is someone who has taken a system from nothing to production and owned it end to end. You picked the technology, argued the tradeoffs in writing, provisioned the infrastructure, and were responsible for it when it broke.

On AI

We use AI heavily and we want you to. What we are not looking for is someone who outsources their judgment to it. The mental model of this system has to live in your head, not in a context window. Use agents to move quickly on implementation. Do the design, the reasoning and the writing yourself, and be able to defend every decision you ship.

What you'll do

Four areas, roughly in the order you will spend time on them. You won't work on all of them at once, but you should be open to any of them.

• Product platform. Make the easiest way for product engineers to do something (the paved road, or golden path) also the most secure, reliable and scalable way by default. Build the shared primitives, libraries and APIs that hide complexity and carry our quality and observability standards with them. Own the architecture and the core stack: what gets standardized, what gets reused, and where the system should be in eighteen months.


• Infrastructure, observability and reliability. Infrastructure as code by default, from cloud resources through to dashboards and alerts. Provision, run and tune our Kubernetes cluster and cloud footprint, and keep CI fast as the deploy rate grows. Define SLIs for the workloads that matter, build SLOs on top of them, and make the alerting high-signal enough that people trust it. Drive down AI and infrastructure spend.


• Security and compliance. Keep the internal foundations secure by default: IAM, dependency management, secrets. Own the authentication and authorization stack, including ReBAC models covering both humans and agents. Implement SOC 2 and ISO 27001 controls without taxing every future change.


• AI dev tooling. Keep product engineers and their agents on the paved road by making sure the documentation and agent guidelines they need are in place. Shorten the path from design to implementation with standardized automations and development environments that stay close to production.

A system of record

A system of record for managing and distributing constantly evolving agent configurations, under strict auditability and tenant isolation requirements.

The harnesses powering our AI agents

Abstracting over use cases that change faster than the code.

Centralized progress tracking

That holds up across a growing number of parallel agent executions.

Usage-based tenant billing

With quotas and rate limiting across a growing set of products.

SLIs for the background job system behind our agents

Execution latency, AI spend, memory, CPU), then fixing the bottlenecks they expose at both the application and infrastructure level.

A document pipeline

Handling dozens of file formats, very large spreadsheets and hundreds of parallel uploads, under a strong reliability requirement.

You will fit into this role if you…

Have designed, built and operated distributed…

Senior/Staff Platform Engineer (m/f/x) — Cortea AI, Berlin.