Hi, I builddelightfulsystems.
AI Engineer at Amsterdam Data Collective, building production agent systems and applied ML products.
Highlighted projects
Selected work showcasing my expertise in building practical, scalable AI systems that deliver real-world value.
An AI-first app for conference attendees, built to serve thousands of simultaneous users.
Backend and ML implementation for ADC's living AI-driven brand identity: embeddings, similarity search, pattern fusion, and preference feedback loops.
Editorial AI workflow for turning technical medical articles into accessible plain-language summaries with grounding, review, and approval built in.
Custom production coding agent and harness with streaming, resumable state, sandboxed file work, MCP tooling, and long-running context management.
Technical skills
What I use to build AI systems end to end, from the agent harness to the infrastructure it runs on.
Agentic systems
Harness engineering for custom agents: tools, skills, MCP connections, long-running context and evaluation loops.
AI infrastructure
IaC and CI/CD, observability and logging, AI gateways and controls, sandboxing and containers.
Cloud and data
Azure, AWS and Google Cloud, plus the structured and unstructured data storage behind each product.
Languages and frameworks
Python for backends and models. TypeScript and React for the interfaces people use.
Career timeline
A compact path through the consulting, engineering, and applied ML work that shaped the current AI engineering focus.
AI Engineer
Amsterdam Data Collective
Building production AI systems across brand tooling, editorial workflows, and custom agent infrastructure.
AI & Machine Learning Consultant
PwC Data Intelligence
Translating business goals into production AI systems across document automation, semantic search, computer vision, and data platforms.
Machine Learning Researcher
Salling Group
Master thesis project on transformer-based recommendation systems for retail customer experience and incremental revenue.
Deep Learning Project
Deloitte / DTU
Built semantic segmentation models for vehicle part recognition in insurance damage assessment.
Business, Philosophy, and AI/ML
Copenhagen Business School & DTU
Business education paired with extensive self-directed and project-based machine learning work.
Get in touch
Reach out for AI engineering work, product collaborations, or technical conversations.