# Writing an AI job posting that attracts the right people [Skip to content](#lm-inhoud)Network/[NL](/en/een-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt)EN[Hubhub.llmnet.nlCompare models on task, language, cost and licence.](https://hub.llmnet.nl/en/)[Communitycommunity.llmnet.nlPrompt techniques, patterns and system prompts.](https://community.llmnet.nl/en/)[APIapi.llmnet.nlLLMs in production: rate limits, routing, structured output.](https://api.llmnet.nl/en/)[Consultancyconsultancy.llmnet.nlRolling out AI in an organisation, pilot to production.](https://consultancy.llmnet.nl/en/)[Newsnieuws.llmnet.nlAI developments, explained for the Netherlands.](https://nieuws.llmnet.nl/en/)[Benchmarkbenchmark.llmnet.nlMeasure AI quality yourself, on your own tasks.](https://benchmark.llmnet.nl/en/)[Careersvacatures.llmnet.nlAI roles, salaries and career paths in the Netherlands.](https://vacatures.llmnet.nl/en/)[Learnleren.llmnet.nlAI concepts in plain language, beginner to builder.](https://leren.llmnet.nl/en/)[Guidegids.llmnet.nlRun AI privately on your own Mac, PC, NAS or home server.](https://gids.llmnet.nl/en/)[Directorydirectory.llmnet.nlMapping the AI ecosystem: tools, models, companies.](https://directory.llmnet.nl/en/)[Radarradar.llmnet.nlSignals from X, research and communities for indie developers.](https://radar.llmnet.nl/en/)[Appsapps.llmnet.nlReviews of AI apps and open-source repos, with tips for builders.](https://apps.llmnet.nl/en/)[llmnet.nl — main site](https://llmnet.nl/en/)[](https://x.com/intent/post?url=https%3A%2F%2Fvacatures.llmnet.nl%2Fen%2Feen-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt&text=Writing%20an%20AI%20job%20posting%20that%20attracts%20the%20right%20people)[](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fvacatures.llmnet.nl%2Fen%2Feen-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt)[](https://www.reddit.com/submit?url=https%3A%2F%2Fvacatures.llmnet.nl%2Fen%2Feen-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt&title=Writing%20an%20AI%20job%20posting%20that%20attracts%20the%20right%20people)[](#)[](https://x.com/intent/post?url=https%3A%2F%2Fvacatures.llmnet.nl%2Fen%2Feen-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt&text=Writing%20an%20AI%20job%20posting%20that%20attracts%20the%20right%20people)[](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fvacatures.llmnet.nl%2Fen%2Feen-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt)[](https://www.reddit.com/submit?url=https%3A%2F%2Fvacatures.llmnet.nl%2Fen%2Feen-ai-vacaturetekst-schrijven-die-de-juiste-mensen-aantrekt&title=Writing%20an%20AI%20job%20posting%20that%20attracts%20the%20right%20people)[](#) # Writing an AI job posting that attracts the right people By Ivo Donker — compiled with AI assistance (Claude & Gemini) · Last updated: August 7, 2026 -> This article is written for employers, HR professionals, hiring managers, and tech leads who are responsible for recruiting and selecting AI specialists in the Dutch labor market. Writing a job posting for an artificial intelligence role requires a fundamentally different approach than recruitment for traditional IT or data roles. The Dutch labor market for AI professionals is characterized by persistent scarcity, a high degree of role fluidity, and a pool of candidates who look very critically at an organization's actual technical maturity. To understand how this scarcity shapes the dynamics in the workplace, it's advisable to consult the analysis on the [Dutch AI labor market and market dynamics](https://vacatures.llmnet.nl/en/ai-arbeidsmarkt-nl) . In today's market, a job posting is no longer a passive recruitment tool but an initial substantive filter. A poorly worded job posting not only leads to a lack of qualified responses, but can also damage the employer brand among experienced specialists. Many organizations struggle to balance attracting sufficient volume with selecting for the right competencies. The process of drafting a sharp job posting takes initial time from the engineering team and forces the organization to make choices in advance about the actual work involved, which can be experienced as an organizational drawback. However, this upfront investment is necessary to prevent miscasting and long-standing open vacancies. ## The two dominant failure patterns in AI job postings In practice, Dutch AI job postings often show two extreme failure patterns that both result in an inefficient recruitment process. The first failure pattern is the overly promotional but substantively empty job posting. These texts are full of generic qualifications such as 'passionate pioneer' or 'data enthusiast', but offer no concrete information about the existing data structure, the available infrastructure, or the specific problem the new employee needs to solve. The direct consequence of this is a high influx of applicants with insufficient technical grounding. Processing these large numbers of applications puts heavy pressure on the HR department, which spends its time rejecting unsuitable candidates. The second failure pattern is the encyclopedic wish list. Here, the employer demands experience with a pile-up of frameworks, platforms, and methodologies that are rarely mastered by a single professional in practice. Job postings that require simultaneous in-depth knowledge of model architecture, MLOps pipelines, cloud infrastructure, and specific business systems actually scare off skilled candidates. Experienced specialists know that such profiles realistically don't exist, or that the role will in practice come down to putting out organizational fires. The drawback of this failure pattern is that the organization is confronted with zero quality responses, leaving the vacancy open for months without the internal hiring need being fulfilled. Failure pattern | Characteristics in the text | Direct effect on recruitment | Organizational drawback | Substantively empty recruitment text | Lots of adjectives, no infrastructural details. | High influx of unqualified applicants. | Time loss for HR and high rejection rates. | Encyclopedic wish list | Long lists of tools, frameworks, and impossible experience requirements. | Drop-off of experienced generalists and specialists. | Zero influx, long-standing open vacancies. | ## First sharpen the focus on the role and the problem Before a single word of the job posting is written, the organization must determine what concrete problem the role needs to solve. Many organizations publish a vacancy for an 'AI Engineer' because it's a popular market term, while the actual problem calls for a data engineer, a traditional software developer, or a technical product manager. An incorrect role definition leads to a mismatch in expectations during the probationary period. Organizing roles within a salary structure and defining the boundaries of tasks happens through the job architecture (functiehuis); consult the guide on [the Dutch AI job architecture and role definitions](https://vacatures.llmnet.nl/en/ai-functies-uitgelegd) to determine the right job title. By making the job title match the desired responsibilities exactly, the employer prevents applicants from responding based on an incorrect picture of the day-to-day work. A drawback of a very specific role definition is that the target audience becomes smaller, but the relevance of the responses increases proportionally. For organizations that don't yet have a clear picture of how a new AI role relates to existing work processes, an analysis of the change in tasks is necessary. Mapping out in advance which tasks within an organization will change can be done via the analysis of [impact assessments on roles and tasks](https://consultancy.llmnet.nl/en/ai-impactassessment-functies-en-taken), which prevents an unrealistic profile from being drawn up. Without this prior analysis, there is a risk that a job posting is drafted based on wishful thinking from management rather than the actual operational need. ## Level, seniority, and separating requirements from preferences Determining the required seniority level in AI roles requires a nuanced approach. In a field that develops rapidly, setting requirements such as 'five years of experience with generative models' is practically impossible or indicates a lack of market knowledge on the employer's part. Seniority in AI roles is not primarily measured by the number of years a specific technique has existed, but by general software engineering principles, system architecture, data discipline, and the ability to maintain complex models in production. In the job posting, a strict distinction must be made between hard requirements and desired experience. Hard requirements are the skills without which an employee gets stuck within the first month. Desired experience includes technologies the candidate can learn within the organization. When an employer classifies too many requirements as 'hard', they exclude valuable candidates who have strong foundational skills but haven't yet applied a specific framework. A clear division of roles and skills prevents technical and operational responsibilities from getting mixed up; see the overview of [AI roles and skills in a non-technical context](https://vacatures.llmnet.nl/en/ai-vaardigheden-in-niet-technische-functies) to determine which digital skills are required for supporting and multidisciplinary roles. Explicitly stating the expected level of independence per task gives applicants a realistic picture of the amount of guidance available. ## Setting requirements for built work and portfolios One of the most important pillars of an effective AI job posting is shifting the focus from certificates to built work. Formal certifications from cloud providers or short online courses offer little guarantee of the ability to write robust code or set up complex data models. The job posting should therefore explicitly ask for evidence of practical execution. A portfolio comprises the built work and codebase with which an applicant demonstrates practical competence; read the article on [assessing résumés and portfolios for AI roles](https://vacatures.llmnet.nl/en/cv-schrijven-voor-ai-functies) to see how candidates present their body of work and how it can be assessed. The job posting could, for example, ask for a link to a public repository, a contribution to open-source projects, or a description of a system taken into production, including the candidate's own specific role in it. Point of attention for the employer: Including a portfolio requirement in the job posting raises the barrier for applicants. This reduces the total number of responses but increases the quality. However, the organization must ensure that internal engineering capacity is available to carefully assess these portfolios in substance. When an organization doesn't have the capacity to assess portfolios in substance, setting this requirement in the job posting is pointless. Applicants quickly notice during the first conversation whether their submitted work was actually reviewed, and failing to do so undermines the credibility of the selection process. ## Dutch labor market context: CAO, salary, and conditions The Dutch labor market has specific characteristics that need to be communicated clearly in the job posting. Applicants in the AI domain often compare offers from Dutch employers with international or remote-working organizations. Withholding the terms of employment or using vague terms like 'market-conform salary' leads to distrust among experienced candidates and a lower conversion rate from vacancy views to applications. Salary forms the primary financial compensation for the role and is often tied to cao pay scales or market-conform ranges in the Netherlands; see the analysis on [the societal and administrative developments surrounding AI and work](https://nieuws.llmnet.nl/en/ai-en-werk) for the broader context of the Dutch polder model. When a role falls under a collective labor agreement (CAO), this must be explicitly stated, including the relevant pay scale and the growth opportunities within that scale. In addition to the gross monthly salary, secondary employment benefits are decisive in the Netherlands. The job posting should concretely mention the following elements: - Pension scheme: The employer's pension contribution and the type of pension fund or insurer. - Holiday allowance and extra payments: The holiday allowance percentage (standard 8%) and any fixed thirteenth month or performance bonuses. - Travel expense reimbursement and work-from-home policy: The reimbursement per kilometer or for public transport, the available work-from-home budget, and the mandatory office days per week. - Hardware and training budget: The specific equipment provided and the annual budget for personal development. The drawback of transparently stating salary scales upfront is that it limits the employer's negotiating room and exposes any internal pay differences with existing staff. Still, the gain in efficiency and trust among suitable candidates far outweighs this. ## The selection procedure as an integral part of the job posting A skilled AI professional wants to know in advance what the selection process looks like and how much time it will take. A job posting that closes with a transparent step-by-step plan stands out positively in a market where lengthy and unclear procedures are common. The description of the procedure should concretely state how many rounds the process consists of, who will be at the table, and what the turnaround time is. To understand how applicants dissect job postings and filter for inconsistencies or red flags, it's advisable to consult the guide for [critically reading AI job postings](https://vacatures.llmnet.nl/en/ai-vacatureteksten-lezen) . By holding up the applicant's mirror, the employer learns which phrasings unintentionally raise doubts about the internal organization. Employers who still need to shape their department structure or selection committee can use the guidelines for [building and structuring an AI team](https://vacatures.llmnet.nl/en/ai-team-samenstellen) to ensure the right technical disciplines are represented during the interviews. After all, an applicant for a senior engineering role expects to speak with a substantive peer, not exclusively with HR generalists. If a technical assessment is part of the process, the job posting must be clear about this. To effectively structure the technical assessment and align it with the market, the guide on [preparing a technical assessment](https://vacatures.llmnet.nl/en/technisch-assessment-voorbereiden) offers practical starting points for both the content and the expected time commitment. An unpaid take-home assignment that takes more than four hours leads to a high drop-off rate in today's market. A shorter, focused case or a live pairing session works better in practice. ## Aligning with the organization type and recruitment channels The tone and content of the job posting must closely match the reality of the organization type. A startup that promises the latest infrastructure but actually has an outdated legacy environment will quickly lose new employees. Likewise, a corporate organization must be honest about the governance and privacy procedures involved in developing AI systems. Depending on the type of organization, the substantive appeal of the text differs, as detailed in the comparative analysis on [working at AI startups versus corporates](https://vacatures.llmnet.nl/en/werken-bij-ai-startups-vs-corporates). Where a startup emphasizes autonomy and direct influence on the product, a corporate will instead emphasize the scalability of data flows, budgets for cloud infrastructure, and the stability of the work environment. For the final distribution of the job posting, simply placing it on general job boards is rarely sufficient. For targeted distribution off the beaten path, we refer to the overview of [networks and physical meeting places in the Dutch AI sector](https://vacatures.llmnet.nl/en/netwerken-in-de-nederlandse-ai-sector) where employers and candidates meet each other on substantive grounds. ## What definitely does NOT work in an AI job posting To ensure that a job posting doesn't send the wrong signals, employers should explicitly avoid the following elements: - Piling up jargon without context: Merely naming popular terms without explaining how they are applied in daily practice. - Vague promises about data quality: Claiming that all data is 'clean and ready to use immediately', while in practice a large part of the work consists of cleaning and structuring. Being honest about this builds trust. - No distinction between research and production: Presenting a role as an R&D position focused on building new models, while the actual task is maintaining and connecting existing APIs. - Unrealistic requirements for junior roles: Asking for several years of production experience for an entry-level role or traineeship. - The absence of contact persons with subject-matter knowledge: Referring exclusively to a general HR email address without offering the opportunity to ask questions in advance to a team lead or senior developer. ## Step-by-step structure of an effective AI job posting A well-structured job posting for an AI role follows a logical and transparent structure that guides the candidate step by step through the essence of the role. By using the step-by-step plan below, the employer ensures that all critical elements are covered without lapsing into marketing language or unrealistic lists of requirements. - The core of the role (first paragraph): Describe in a maximum of three sentences what business or technical problem this role solves, which team the candidate will work in, and what the direct impact of the work is. - The technical context and environment: Give a concrete overview of the current tech stack, the cloud infrastructure used, the data flows, and the extent to which models are already running in production. - Responsibilities and daily tasks: Describe four to six concrete tasks, making clear the ratio between building, maintaining, consulting, and documenting. - Required profile and hard requirements: Formulate a maximum of four to five hard requirements, focused on demonstrable skills and built work rather than years of experience or certificates. - Employment terms and CAO: State the salary indication or scale, the pension scheme, the training budget, and the hybrid work arrangements. - The selection process: Provide a step-by-step overview of the number of interviews, any assessment, and the expected total turnaround time. Writing an AI job posting according to these principles requires more upfront alignment between HR, management, and the technical teams. The drawback is that drafting the text takes more time than copying a standard template. The benefit, however, is a significantly higher relevance of applicants, a shorter turnaround time for the selection procedure, and a more durable match in the workplace. © 2026 LLMnet.nl Vacatures. All rights reserved.