Agent Developer
Wilmo.ai
Agent
Developer
Eval-driven development. Three production agents. Synthetic worlds so they can refund, update orders, and send email without touching live systems.
The Role
Building agents starts with great evals
Wilmo already has multiple agents in production, solving tickets autonomously. We are at the edge of letting agents execute tasks end-to-end on behalf of customers — safely.
Those agents lean on more than 60 integrations. They need to read and write. To keep production systems intact, we built internal worlds of synthetic data: sandboxed harnesses where an agent can update an order, issue a refund, or send an email.
Your job is to keep building the customer-facing agents. Today they are all reactive. The interesting work is making them proactive.
Day to day
Start with the eval
We build agents with Eval Driven Development. If you cannot measure it, you do not ship it.
Improve the three production agents
Wilmo, Wilbot, and the Chatbot. Each has its own evals. Each is already live, resolving real tickets.
Keep the worlds honest
Agents read and write across 60+ integrations. Synthetic worlds let them refund, update orders, and send email without touching production.
Move from reactive to proactive
Every agent we have waits for a ticket. The next problem is agents that act before the customer writes.
01
Wilmo
Resolves tickets asynchronously, end-to-end.
02
Wilbot
Internal CS agent that trains and improves Wilmo.
03
Chatbot
Website chatbot, agentified — it solves, it does not link.
About You
You'll thrive here if
You think in evals
You do not prompt until it looks good. You write the test that tells you whether the agent actually got better.
You respect production
Refunds, order updates, and outbound email are not toys. You want harnesses that make it safe to be ambitious.
You like hard problem spaces
This is not a wrapper around an API. It is agents that execute real work for real shops.
You want to learn from the person who built it
You will sit next to Villads every day and learn how the agentic infrastructure was built from the ground.
Compensation & Benefits
50,000 DKK / month, growing with MRR
+2,500 DKK for every 100k MRR the company adds, for the next 500k MRR.
Today50,000DKK
+100k MRR52,500DKK
+200k MRR55,000DKK
+300k MRR57,500DKK
+400k MRR60,000DKK
+500k MRR62,500DKK
50k
Today
52.5k
+100k
55k
+200k
57.5k
+300k
60k
+400k
62.5k
+500k
0.5% warrants
Own a real piece of the company you help build.
Copenhagen office
Work with the team in person from our office in Indre By. This is a full-time, office-based role.
Meals covered
Breakfast, lunch, dinner, and snacks are covered when you are in the office.
This Is Not
- ✕A 9 to 5 job
- ✕A place to prompt until the demo looks good
- ✕A research lab with no production agents
Sounds like you?
No CV needed. Just a short video, your LinkedIn, and one question.
Full Name
2-Minute Loom Video
Record a short video introducing yourself and why you want this role. Free at loom.com
LinkedIn Profile
Tell us about a time you evaluated or improved an AI system. What did you measure, and what changed?
Would you be able to work on-site at our Copenhagen office?
YesNo