I am Caspar Bannink, an AI engineer, full-stack software engineer, and founder with a background spanning both AI research and production software.

I hold both a BSc and MSc in Artificial Intelligence and have spent the past three years working professionally as an AI engineer, alongside more than six years of experience in full-stack software engineering. My work sits at the intersection of AI models and the systems required to make them genuinely useful: agents, LLM infrastructure, integrations, production workflows, local AI, and the economics of deploying these systems.

I am also the founder of HomeScout, an AI-powered rental workflow product. Building a real product keeps me interested in the less glamorous parts of AI that often matter most in practice: reliability, cost, latency, tooling, integrations, and whether a promising model actually improves the end product.

On Medium, I write about the systems behind modern AI rather than simply repeating release announcements. I am particularly interested in coding agents, model and harness design, open and local AI, production AI, infrastructure, and the practical decisions developers have to make as the ecosystem changes.

I originally started writing because I enjoy breaking down difficult technical concepts and making them easier to understand. That is still the goal, but the focus has expanded from explaining AI fundamentals to investigating how today's AI systems actually work, where the hype outruns the evidence, and what is useful for builders.

You can also find me on LinkedIn.