// projects
From visual learning experiments to building an application around an AI agent: these projects show what I’ve built, the decisions behind it, and what I’m learning along the way.
Lessons from building an agentic AI application with LangGraph, durable state, access control and observability, and preparing it for a first release.
A Python and Unity pipeline for vision-based reinforcement learning, exploring how Slow Feature Analysis and 32 learned visual features affect navigation.
Detecting fire safety equipment in escape plans with Keypoint R-CNN and custom annotations, connecting displaced symbols with actual equipment locations.
// say hello
A question about a project, an idea, or just a shared interest in AI. Feel free to drop me a line.
laurenz.hoener@gmail.com