The demo of an agent is easy. The version you can trust with a customer’s money is not. This team has the first and needs someone who can build the second.
What you will do
- Build agents that use tools well and fail safely when they cannot.
- Design the orchestration so the system is debuggable when it breaks.
- Decide what the model should do and what plain code should do.
- Measure reliability and drive it up, week over week.
What we look for
- You have shipped an agent that real people depended on.
- You know the failure modes of tool-use and design around them.
- You are suspicious of demos, including your own.
What stays open
This is posted as a role here, but how you work with the company is something we sort out together. Apply and we will talk.
More about the applied ai engineer track →