Agent Engineer

Agent Engineer

Build something that matters. Right now, over 5,000 companies, from DAX enterprises like Merck to fast-growing startups, use Langdock every day. Their employees open our product to draft strategies, analyze documents, automate workflows, and think through hard problems with AI.

We are the platform that makes this possible: secure, model agnostic, and deeply integrated with how work actually happens.

What "Agent Engineer" Means Here

AI agents are becoming the operating system of how companies work. Not chatbots. Not simple automations. Agents that own tasks end-to-end: reading customer tickets, updating CRMs, delegating subtasks to other agents, and knowing when to escalate to a human.

At Langdock, we are building both the product that lets our customers deploy agents and the internal infrastructure that runs our own. This role sits at that intersection. You will design, build, and operate the AI agents that power critical parts of Langdock's operations, from customer support automation to internal workflows. And the patterns you develop will directly inform how we build the agent platform for thousands of other companies.

This is not a role where you configure tools built by someone else. You will work across the full stack: prompt engineering, orchestration protocols, integrations, reliability, and cost management. You will own agents the way an engineer owns a production service.

What You Will Actually Do

• Own production agents end-to-end. You design, deploy, monitor, and improve the AI agents that handle real customer interactions and internal workflows. When an agent misbehaves, you dig into the logs to identify the root cause and fix it. When it works beautifully, that is also you.

• Build and operate customer support automation. You will monitor Langdock's support experience by supervising agents that resolve issues faster than most humans could, while knowing exactly when to hand off.

• Manage agent reliability, cost, and quality. You track how agents perform, what they cost, where they fail, and why. You set budgets, tune behavior, adjust governance rules, and make sure every agent earns its keep.

• Design the orchestration layer. You define how agent teams are structured, set up approval gates, and build the operational infrastructure that lets autonomous agents work safely at scale.

• Integrate agents with systems. You connect agents to both internal and external APIs, giving them the context and capabilities they need to do real work.

• Experiment constantly. New models drop, new techniques emerge, new use cases surface. You are the person who tries them first, benchmarks them honestly, and ships the ones that actually improve outcomes.

What Makes This Different

This role barely existed two years ago. There is no playbook for it. You are building the discipline of "agent operations" at a company that is also building the platform for it.

The agents you run will directly reach Langdock's 5,000+ customers, meaning your work has an immediate, measurable impact. And because the patterns you develop internally become the patterns we productize for customers, you are not just operating agents. You are defining how agents should be operated.

We went from 2 to 20M+ ARR in one year. You will join early enough to shape how things work, but late enough that we have traction, customers, and a product people love.

You will also learn fast. Our team is small, the scope is large, and the feedback loops are short. People who joined a year ago are now running critical functions. If you are good, you will grow.

You Might Be a Fit If...

• You have built and maintained AI agents or automations in production.

• You are deeply technical. You understand LLMs, token economics, prompt engineering, API design, and orchestration patterns. You can read a protocol spec and immediately see the edge cases.

• You obsess over experimentation and reliability. You think in terms of error rates and customer satisfaction. You instrument everything.

• You are not precious about your agents. When one fails, you find the root cause and fix it. When a simpler approach works better, you kill the clever one.

• You do not just use AI tools daily. You actively nerd about your setup. You experiment with different models, prompts, workflows, and automations. You have strong opinions about what works and why.

• You would rather own a problem than be told exactly what to do.

• You are a kind person who cares about the people around you.