Agentic AI Engineer

Role overview

A senior engineering role dedicated to the difficult middle ground of agentic AI: turning proof-of-concept agents into systems that survive real users, real load, and real failure modes. The position spans two complementary tracks, building multi-agent pipelines and Claude/MCP tooling for client projects while shaping the internal standards, evaluation practices, and guardrails the broader team will rely on. The work targets problems the field has not yet solved cleanly, including context management, regression catching, and oversight loops that actually add safety rather than latency.

Responsibilities

- Architect and ship multi-agent orchestration systems, including the surrounding infrastructure that keeps them dependable past the demo stage - Develop Claude skills, Model Context Protocol integrations, and developer tooling that the rest of the engineering organization uses day-to-day - Mature agent prototypes into production services, addressing observability, retry behavior, and graceful degradation - Establish evaluation frameworks, guardrail patterns, and review gates that define what "safe to ship" means across the team - Feed lessons from production client work back into the internal toolchain

Requirements

- Five or more years in a senior engineering role with hands-on responsibility for shipping systems to production rather than to staging - Substantial experience building with large language model APIs, not just surface-level chat integrations - Strong grasp of agent architecture fundamentals: tool use, multi-agent coordination, context window strategy, and the failure modes that rarely make it into documentation - Demonstrated ability to design around model limits through evaluations, guardrails, human-in-the-loop controls, and realistic scope-setting - Working proficiency in TypeScript and the broader Node.js ecosystem

Nice to have

- Background shipping multi-agent systems in regulated or high-consequence environments - Experience writing Claude Code skills, MCP servers, or custom tooling for coding agents - Open-source contributions in the agent tooling or AI infrastructure space

Benefits and work setup

- Remote-first setup available across Europe - Senior peer group and a culture where the most interesting problems are not reserved by rank - Direct influence on the engineering roadmap for bringing agentic AI to production - Opportunity to help shape how the team builds with agents while that playbook is still being written