Implementation Engineer, AI Agents

AI AGENT ENGINEER

Department: AI Delivery Location: Copenhagen Working model: Onsite Reports to: AI Product Lead Salary: Competitive

ABOUT AUTOFLOWS

Autoflows is an agentic after-sales platform for automotive retail. We deploy AI agents across dealerships in Europe that handle customer conversations end to end across voice and messaging. Our agents identify the right moment to contact customers about service, MOT, warranty and other after-sales opportunities, then handle the conversation, book the appointment and write the outcome back into the dealership's systems. Our platform connects with DMS, CRM and OEM data, turning after-sales work that traditionally sits in queues into work that runs itself. We are now scaling across clients, markets and languages. That requires us to build a delivery capability that does not depend on one person knowing how everything works.

THE ROLE

As an AI Agent Engineer, you will build, deploy, debug and continuously improve the conversational agents running across our clients. You will work across the full lifecycle: Build → Test → Deploy → Monitor → Debug → Improve You will start closely with the AI Product Lead and progressively take independent ownership of production agents and client implementations. The role is not just about configuring agents. With time you will also help build the tooling and systems that make agents easier to deploy, operate and maintain at scale. Architectural decisions for the core agentic platform stay with the AI Product Lead; your platform involvement is about the tooling layer around agents, not the underlying architecture itself.

WHAT YOU'LL DO

Agent development

• Build and tune agents across voice, chat, SMS and WhatsApp
• Develop prompts, tools, workflows, guardrails and fallback paths
• Test agents against realistic customer scenarios
• Improve agent behavior using production data and conversation analysis

Production & implementation

• Diagnose and fix live issues such as model failures, mispronunciations, failed tool calls, broken workflows and missing fallback paths
• Own new agent deployments as dealership clients onboard
• Adapt agents to different processes, languages and markets
• Work directly with clients when technical input is required
• Identify root causes rather than applying temporary fixes

Integrations

• Work with DMS, CRM and other dealership systems
• Configure and troubleshoot integrations for customer data and workshop bookings
• Build reusable patterns for recurring integration requirements

Agentic platform tooling

• We are building an agentic platform, not a collection of manually maintained bots. You will be encouraged to identify repetitive work and turn it into systems. This may include:
• Automated testing and evaluation
• Monitoring and regression detection
• Automated diagnosis of recurring failures
• Self-healing or automated recovery for well-understood failure modes
• Automated deployment and configuration
• Reusable agent components and workflows
• Agent-assisted documentation and operational tooling
• Systems that allow agents to improve from production feedback within defined quality and safety boundaries

| "The principle is simple: if we solve the same problem repeatedly, we should build a system that solves it for us." Architecture & delivery

• Work with the AI Product Lead on complex architectural decisions
• Help determine when to use prompting, workflow logic, deterministic code, specialised agents or other approaches
• Document recurring solutions and contribute to the team's technical playbook
• Help establish standards that make future deployments faster and more reliable

WHAT WE'RE LOOKING FOR

This role sits between AI, technology and human behaviour. Technical ability is necessary, but technical ability alone is not enough. Our agents speak to real customers. You need to care about what the interaction feels like, not just whether the system technically works. Technical

• Hands-on experience building with LLMs and AI APIs
• Strong understanding of prompting and model behaviour
• Experience debugging AI or software systems
• Ability to reason through ambiguous technical problems
• Comfort working without a finished specification
• Ability to learn unfamiliar platforms and systems quickly

Human & social

• Strong social skills and clear communication
• Natural curiosity about how people think, speak and respond
• Ability to look at an interaction from the customer's perspective
• Good judgement about tone, context and conversational nuance
• Comfortable speaking directly with clients and asking the questions needed to understand a problem
• Able to explain technical issues clearly to non-technical people
• Open to feedback and willing to change your approach when evidence shows it is not working
• Interested in how language and culture affect conversational AI

Builder mindset

• You investigate why something failed rather than simply fixing the visible symptom
• You notice repetitive work and look for ways to automate it
• You are comfortable experimenting, testing and iterating
• You think beyond the individual agent and look for ways to improve the underlying platform

Strong plus

• Conversational AI or agent-building experience
• Voice AI, chatbots or workflow automation
• API and systems integration experience
• Experience with tools, function calling or agentic workflows
• Experience operating production AI systems
• Experience working across multiple languages
• Experience with CRM, ERP, DMS or similar business systems
• Automotive experience is not required.

| "We would rather hire a technically capable builder who understands people than a technically exceptional engineer who does not."

SUCCESS IN THE ROLE

As you become fully productive, you should be able to:

• Independently diagnose and resolve most common agent failures
• Build and deploy agents with limited supervision
• Own production improvements from investigation through deployment
• Handle recurring integration problems
• Turn repeated manual work into reusable systems
• Improve the reliability and scalability of our agent platform
• Reduce the amount of hands-on implementation required from the AI Product Lead

THE TEAM

You will work in a small AI Delivery team: AI Product Lead - Owns product direction, architecture, priorities, technical standards and complex technical decisions AI Agent Engineer - Owns agent development, implementation, production debugging and platform improvements AI/ML Student - Supports testing, evaluation, research, data work, experimentation and implementation Because the team is small, you will have significant ownership and direct exposure to the systems we are building.

HOW TO APPLY

Send your CV along with: Show us something you built, broke or fixed. A link, repository, write-up, video or before/after example. We care more about how you approached the problem than how polished the result is. Tell us why this problem interests you. Give us a short note explaining what attracts you to conversational AI, agentic systems and the challenge of making them work reliably in the real world. Contact: Maz Katinas, AI Product Lead