Services
AI, automation
and the software around it.
I focus on AI engineering and automation, and I build the full stack they need to work: web platforms, apps, CMS/CRM integrations, APIs and databases. One engineer, one accountable point of contact.
AI Engineering
I build AI into real software: agents that call your APIs, retrieval (RAG) over your own documents, and LLM features that are tested, observable and affordable to run. Models are chosen per task and can be swapped later, so you're not locked in.
What you get
- AI agents with tool use that act on your systems
- RAG search and Q&A over documents, wikis and tickets
- Model selection across OpenAI, Anthropic, Mistral and open-weight models
- Evaluation sets, guardrails and fallbacks before go-live
- Usage, latency and cost monitoring in production
Typical use cases
- A support copilot that drafts answers from your knowledge base
- An internal assistant that queries your database in plain language
- Classifying and routing incoming requests automatically
AI Automation
Most manual work is reading something, deciding and typing it somewhere else. AI automation takes on that loop: it extracts data from emails, PDFs and forms, checks it against your rules, and writes it into your systems, with a human review step wherever a mistake would be costly.
What you get
- Document and email intake with structured data extraction
- Human-in-the-loop review screens for uncertain cases
- Integrations with CRM, ERP, accounting and ticketing tools
- Audit trails of every automated decision
- Error handling, retries and alerts
Typical use cases
- Invoices and receipts booked straight into your accounting software
- Incoming leads enriched, scored and assigned
- Customer emails sorted, summarised and routed to the right team
Process Automation
Not every problem needs AI. I start by mapping how the work actually flows today, remove the steps that shouldn't exist, then automate the rest with the simplest reliable tool: a scheduled job, an API integration, n8n or Make, or a small custom service.
What you get
- Process mapping and automation roadmap
- System-to-system integrations and data sync
- Scheduled jobs, reports and notifications
- No-code / low-code flows where they're the right fit
- Documentation your team can own
Typical use cases
- Syncing orders between webshop, warehouse and bookkeeping
- Weekly reports generated and emailed automatically
- Onboarding steps triggered from a single form
Prompt Engineering & LLM Ops
An impressive demo prompt and a production prompt are different things. I design prompts and schemas for structured, reliable output, build evaluation sets from real examples, and set up regression tests so you can change models or prompts without guessing.
What you get
- System prompts and structured-output (JSON) schemas
- Evaluation datasets and automated scoring
- Prompt versioning and regression testing
- Latency and token-cost optimisation, including caching
- Team workshops on using LLMs effectively
Typical use cases
- Cutting token spend of an existing AI feature
- Making extraction output consistent enough to automate
- Choosing between models with evidence rather than opinion
Web Platforms
From marketing sites to customer portals and SaaS dashboards: I build the whole stack, from interface to database, with performance, accessibility and search visibility treated as requirements, not extras.
What you get
- Web apps, portals and dashboards
- Multilingual websites (e.g. EN / NL) with technical SEO
- Authentication, roles and payments
- Core Web Vitals and accessibility (WCAG) work
- CI/CD, hosting and monitoring set-up
Typical use cases
- A customer portal for orders, documents and support
- An internal tool replacing a tangle of spreadsheets
- A fast, bilingual company website editors can manage
Desktop & Mobile Apps
When a browser isn't enough (offline use, device hardware, a native feel), I build cross-platform apps with React Native, Electron or Tauri, sharing logic with your web platform so there's one system to maintain instead of three.
What you get
- iOS and Android apps with React Native / Expo
- Desktop apps for Windows, macOS and Linux
- Offline-first data sync
- Push notifications and device integrations
- App Store and Play Store release support
Typical use cases
- A field-service app that works without signal
- A desktop tool that automates local files and hardware
- A companion app for an existing web platform
CMS & CRM
Your team should be able to publish content and find customer information without asking a developer. I set up headless CMSs like Payload and Strapi, and connect CRMs like HubSpot, Pipedrive and Salesforce to the rest of your stack.
What you get
- Headless CMS implementation and content modelling
- Editor-friendly previews, workflows and roles
- CRM set-up, clean-up and data migration
- Two-way sync between CRM, website and tools
- Custom fields, pipelines and automations
Typical use cases
- Website leads landing in the CRM with full context
- Moving from a legacy CMS without losing SEO
- One customer record across sales, support and billing
APIs, Databases & Data
Every product above rests on this layer. I design REST and GraphQL APIs, model PostgreSQL and MySQL databases for growth, build data pipelines, and deliver business intelligence dashboards people actually use.
What you get
- REST and GraphQL API design and implementation
- Database modelling, migrations and performance tuning
- Third-party API integrations and webhooks
- ETL pipelines and data warehousing
- BI dashboards in Metabase, Power BI or Tableau
Typical use cases
- A public API so partners can integrate with you
- Speeding up slow queries behind a growing product
- One KPI dashboard combining sales, ops and finance
Ways to work together
Pick the engagement that fits the work.
Fixed-scope project
A clearly defined deliverable with an agreed price and timeline. Best for new products, automations and integrations.
- Written scope and acceptance criteria
- Fixed price, milestone invoicing
- Handover and documentation included
Sprint retainer
A reserved block of time each month for continuous development. Best for evolving products and AI features.
- Monthly planning with priorities you set
- Weekly demos and progress updates
- Flexible scope within the budget
Consulting & audits
Hourly advice, code or architecture reviews, and AI readiness assessments. Best before you commit to a build.
- AI feasibility and cost assessment
- Architecture and code reviews
- Workshops for your team
FAQ
Common questions
Do you work with companies outside Amsterdam?
Yes. I'm based in Amsterdam and work with clients across the Netherlands and the EU, mostly remotely, with on-site sessions when they help.
Which AI models do you use?
Whatever fits the task: OpenAI, Anthropic Claude, Mistral, Google Gemini or open-weight models hosted in the EU. I design systems so the model can be swapped without rewriting the application.
How do you handle privacy and GDPR with AI?
I prefer EU data residency, providers that don't train on your data, minimal data sent to models, and processor agreements in place. Sensitive workloads can run on self-hosted models.
How much does a project cost?
It depends on scope. After a free intro call you get a written proposal with a fixed price or a monthly budget, so there are no open-ended surprises.
Who owns the code?
You do. Repositories, cloud accounts and API keys are created in your name, and the intellectual property of client deliverables transfers to you on payment.
Can you work with my existing team or agency?
Absolutely. I regularly slot in alongside in-house developers, designers or agencies to deliver the AI or automation part of a larger project.
Not sure which service you need?
Describe the problem, not the solution. I'll suggest the simplest thing that works.