Forward deployed AI engineer

Most organizations have done something with AI: a pilot, a demo, a report. And on Monday everyone goes back to working as usual. We are not satisfied with that. At Gaide, we build human-centered AI: systems of collaborating AI agents that carry out complex processes from start to finish, designed around the people who work with them. For us, it is not about the technology, but about what it delivers for the customer: financially, in quality, or in safety.

• You work on-site at customers and turn a vague process problem into an agent, integration, or tool that runs in production.
• You own the architecture: the choice between an agent, workflow, or ordinary code; the structure of context and retrieval; tool use; fallback logic; and integration with existing systems.
• You determine how we evaluate.
• You give final technical approval for production. In doing so, you assess not only the technical aspects, but also the AI System Card and compliance with the AI Act.
• You establish engineering standards and coach engineers on them: code reviews, CI/CD, security, monitoring, and documentation.
• You join meetings with the customer's management and IT teams and translate technical choices into costs, risks, and the next step.
• You help sales define the scope of assignments, estimate the work, and prepare demos.
• You document lessons from practice in reusable building blocks and playbooks, so the next project is faster.
• You lead through subject-matter expertise. You are not a project manager and have no direct reports. You spend about half your time building yourself: writing code, reviewing code, and creating prototypes. The other half goes to architecture, defining the scope of assignments with customers, and helping the team move forward.
• You make the decisions that set this role apart from our Forward Deployed Engineers. You choose the architecture, determine when something is ready for production, and ensure it remains maintainable and affordable after handover.

• You work comfortably without complete specifications and enjoy raising the bar for others.
• You have at least five years of experience as a software, data, ML, or AI engineer. During that time, you have provided technical leadership for projects or a small team.
• You have brought multiple LLM applications (agents, RAG, or workflows) into production, and you were responsible for the architecture of at least one of them.
• You write production Python, have set up CI/CD pipelines yourself, and know one cloud platform in depth: identity, network, and costs, not just the AI services. Most of our customers work with Azure.
• You have built integrations with APIs, databases, and data warehouses, and know how authentication and permissions work in an enterprise environment, for example through OAuth or Entra ID.
• You have set up an evaluation and monitoring approach for an LLM application and used it to support decisions, such as changing models or deciding to go live.
• You know the security risks of LLM applications, such as prompt injection and data leakage through retrieval, and have built measures to address them.
• You can justify an architectural choice in terms of cost, risk, and maintenance, and explain it to a CTO and to the engineers building it.
• You speak fluent Dutch and good English. Most customers and users work in Dutch.
• Будет плюсом: Observability and cost monitoring for GenAI per application, for example with Langfuse or OpenTelemetry; Infrastructure as code, such as Bicep or Terraform; Governance as a design principle: the AI Act, the GDPR, and auditability; MCP or an agent framework such as LangGraph; Work in construction, energy, or infrastructure, or a background in consultancy; A driver's license is useful, because you travel to customer locations

• Salary based on your experience
• A pension plan
• Your own customers from day one, in different sectors
• Influence over how Gaide builds: the standards, building blocks, and playbooks you establish are used by the whole team
• You work together at the office or at customer locations for most of the week. One day per week, you work wherever you want.
• 25 vacation days, full public transport reimbursement, and daily lunch on the Keizersgracht
• Team outings and activities
• A budget for training and conferences