Senior AI Platform Infrastructure Engineer

Luxoft Poland; Polska; On-site; PLN 140,000 - 190,000; Full time

Job summary

Luxoft Poland is seeking Senior Platform Engineers to join a top-tier hedge fund project expanding its AI agent platform. You will own both the agents and MCP tooling and operate the Kubernetes/AWS infrastructure underpinning them.

You'll design secure agent collaboration, robust memory and access controls, and reliable evals at scale, with a focus on multi-tenant, security-sensitive workloads and observable incident response.

Qualifications

• Hands-on production experience building agents and agent-driven systems.
• Experience designing and operating MCP servers and tooling.
• Deep Kubernetes expertise (EKS preferred): cluster design, networking, RBAC and security.

Responsibilities

• Own both layers: agents and MCP servers/tools, and the Kubernetes/AWS platform.
• Design secure agent collaboration and inter-agent communication at scale.
• Build multi-tenant, security-sensitive infrastructures and reliable evals.

Skills

Kubernetes production experience

AWS cloud architecture

CI/CD and GitOps

Infrastructure as code

Security policies & RBAC

Observability & incident response

Multi-tenant security design

Tools

Claude Agent SDK

Codex SDK

Terraform

CloudFormation

Pulumi

Job description

Project Description

Our partner, a top tier Hedge Fund, is expanding their AI agent platform. The platform runs 1,500+ agents in production across the firm. We are looking for Senior Platform Engineers to build across the whole stack, from the infrastructure to the agents and MCP tools that run on it. In the next phase agents securely work together.

Responsibilities

You’ll join the team that runs and own both layers: the agents, MCP servers and tools our users rely on, and the Kubernetes/AWS platform with event-driven systems underneath them. From there, you’ll build what lets agents work as a team: secure communication between agents, agents that manage other agents, shared memory with the right access controls, and the evals that keep it reliable at scale.

Skills required

• Hands-on experience building agents in production with agent SDKs (e.g. Claude Agent SDK, Codex SDK): tool use, context and memory, prompting
• Experience building and running MCP servers and tools
• Deep hands-on Kubernetes (EKS preferred): cluster design, networking (CNI, service mesh), RBAC and security policies
• Strong AWS: VPC architecture, IAM, EKS/ECS, networking and security
• groups Infrastructure as code (Terraform, CloudFormation or Pulumi), GitOps and CI/CD (ArgoCD, GitHub Actions, Jenkins)
• Container architectures for multi-tenant, security-sensitive workloads
• Event-driven systems: queues and pub/sub (e.g. SQS, Kafka), webhooks,
  async workers, retries and idempotency
• Production observability and incident response (Datadog or similar)

What is nice to have

• Agent-to-agent (A2A) protocols and multi-agent orchestration
• UI / front-end experience (a plus for one of the three roles: agent
• Auth: OAuth 2.0 / OIDC, identity federation and SSO, delegated (on-behalf of) tokens
• Permission-aware search, retrieval or shared memory
• Securing autonomous or long-running AI workloads at scale
• Financial or regulated environments, compliance-aware infrastructur