Agentic Engineer

Levy Professionals; Amsterdam; On-site; EUR 90,000 - 120,000; Full time

Job summary

Levy Professionals is assembling a new internal team in the Netherlands to build AI tooling for developers. Our engineers will ship agents that assist writing code, reviewing, testing, and managing backlogs across the bank's dev teams.

The setup is greenfield, with top-level sponsorship and a culture that favours prototypes over slides. Expect fast feedback, room to experiment, and colleagues who focus on delivering working software.

Qualifications

• 5+ years in software engineering, DevOps, ML engineering or data science.
• Software shipped end to end, from design through production and operations.
• Cloud experience plus infrastructure as code (Terraform or Bicep).
• Comfortable across languages and stacks, you pick the tool for the problem.
• Hands-on with coding agents and agentic tooling, in production or serious side projects.

Responsibilities

• Building AI assistants and agents for the software delivery process, from refinement to code generation, review and testing.
• Turning promising prototypes into solid, reusable building blocks other teams can adopt.
• Making the trade-offs on speed, cost, reliability and governance.
• Coaching fellow engineers on getting real value out of coding agents.
• Helping define how the impact of these tools gets measured.

Skills

Software engineering

DevOps

ML engineering

Data science

Job description

One of the larger Dutch banks is standing up a brand new internal team with a single mission: build AI tooling for their own developers. Agents that help write code, review it, test it and keep backlogs in shape. Everything this team ships rolls out to development teams across the entire bank, so a small win here scales hard.

The setup is mostly greenfield. The team is new, sponsored from the very top, and the engineers joining now set the standards instead of inheriting them. Expect short feedback loops, room to experiment and colleagues who'd rather ship a working prototype than write a slide deck.

What you'd be doing:

• Building AI assistants and agents for the software delivery process, from refinement to code generation, review and testing
• Turning promising prototypes into solid, reusable building blocks other teams can adopt
• Making the trade-offs on speed, cost, reliability and governance
• Coaching fellow engineers on getting real value out of coding agents
• Helping define how the impact of these tools gets measured

Who fits: engineers first, GenAI second. The client is explicit about this. Prompting skills alone won't get you through the door. You've been building software since before the current AI wave, your fundamentals are solid, and you also happen to be the person on your team who's deepest into coding agents.

What they ask:

• 5+ years in software engineering, DevOps, ML engineering or data science
• Software shipped end to end, from design through production and operations
• Cloud experience plus infrastructure as code (Terraform or Bicep)
• Comfortable across languages and stacks, you pick the tool for the problem
• Hands-on with coding agents and agentic tooling, in production or in serious side projects
• A hobby build, prototype or repo you can show. Proof of maker energy carries real weight here.
• You live in the Netherlands and have the right to work here. Relocation isn't possible on this one.