> Markdown version of [/videos/1609-from-code-to-motion-building-an-autonomous-hat-hunting-robot-with-kubernetes-ml?t=5](https://www.wearedevelopers.com/videos/1609-from-code-to-motion-building-an-autonomous-hat-hunting-robot-with-kubernetes-ml?t=5). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# From Code to Motion: Building an Autonomous Hat-Hunting Robot with Kubernetes & ML

To solve the nightmare of disconnected edge deployments, engineers built an autonomous, hat-hunting robot. Discover how standard Kubernetes and GitOps push machine learning models from code to physical motion.

- **Speakers:** [Daniel Brintzinger](https://www.wearedevelopers.com/@daniel-brintzinger)
- **Event:** World Congress 2025
- **Published:** August 20, 2025
- **Duration:** 30:58
- **URL:** https://www.wearedevelopers.com/videos/1609-from-code-to-motion-building-an-autonomous-hat-hunting-robot-with-kubernetes-ml

## Summary

"The robot is actually more of a symbol here," representing far-edge devices like disconnected factory floors or remote hardware that suffer from spotty network connectivity. While robust DevOps setups and containerization have resolved lifecycle management in centralized data centers, bringing intelligent workloads to remote infrastructure is notoriously difficult. To demonstrate resolving this gap, the talk walks through building an autonomous, fedora-hunting Raspberry Pi robot. It proves that engineers can leverage standard cloud-native methodologies to push configurations, apps, and machine learning models seamlessly from code ideation to physical motion at the extreme edge.

A central thesis of the presentation is that "AI services are applications in a form by themselves." Because AI inferencing boils down to standard API calls taking inputs and returning outputs, data science workloads can benefit entirely from existing Kubernetes architectures. Inside Red Hat OpenShift, data scientists can spin up Jupyter Lab workbenches, utilize central GPUs to train a YOLOv5 object recognition model on standard image sets, convert the result to the ONNX format, and dump it into an S3 bucket via Elyra pipelines. By treating AI as just another containerized service, teams can reserve heavy compute architecture for model creation while safely deploying the predictive logic downward to minimal hardware footprints.

To orchestrate these centrally authored models onto an unreachable edge node, the physical robot relies on MicroShift, a lightweight Kubernetes distribution. Layer 7 routing via Skupper resolves restrictive local network conditions by securely wrapping TCP packets in AMQP, bridging the core namespaces with the edge device without requiring direct IP access. Once the networking is handled, the robot is pulled into Argo CD and treated as a standard managed cluster. This enables GitOps-driven delivery, ensuring that updates to the object recognition code or application logic automatically roll out to the edge device, unifying distributed development and hardware operation.

**Keywords:** edge deployments, containerized workloads, microshift edge nodes, openshift ai pipelines, skupper network routing, disconnected cluster mapping, gitops edge management, argo cd infrastructure, yolov5 object recognition, onnx model conversion, jupyter lab workbenches, elyra machine learning pipelines, raspberry pi edge computing, ai inferencing apis, devops lifecycle management, tcp packet wrappers, lightweight kubernetes

## Chapters

1. **Challenges of application edge deployment and containers** (00:05) — How DevOps approaches and containers enable application delivery to disconnected edge devices.
1. **Hardware constraints and technology stack for the robot** (02:53) — The physical components and software tools used to build the autonomous robot.
1. **Training an object recognition model with OpenShift AI** (08:59) — Using Jupyter Lab and a machine learning pipeline to train an object detection model.
1. **Deploying and testing the AI inference service** (16:24) — How to expose a trained model as a secure REST API for application consumption.
1. **Building the robot control application using Eclipse Che** (20:19) — Creating a Python application within a cloud-based development environment to interact with the AI service.
1. **Live demonstration of the autonomous hat-hunting robot** (27:30) — Executing the control application to observe the robot scanning for and navigating toward the target.
1. **Continuous deployment to edge devices using Argo CD** (29:41) — Managing the physical robot as a registered Kubernetes cluster to automate configuration and code rollouts.

## Related Moments

- [Speaker background and open source Kubernetes edge computing projects](https://www.wearedevelopers.com/videos/100094-from-bytes-to-execution-writing-a-webassembly-runtime-in-rust) (from "From Bytes to Execution: Writing a WebAssembly Runtime in Rust")
- [Deploying AI-driven foreign object detection in manufacturing environments](https://www.wearedevelopers.com/videos/1415-from-factory-floor-to-kubernetes-core-building-an-edge-platform-one-step-at-a-time) (from "From Factory Floor to Kubernetes Core: Building an Edge Platform One Step at a Time")
- [Navigating hardware constraints and edge computing challenges](https://www.wearedevelopers.com/videos/1452-robots-2-0-when-artificial-intelligence-meets-steel) (from "Robots 2.0: When artificial intelligence meets steel")
- [Deploying parallel models on custom hardware with Edge Impulse](https://www.wearedevelopers.com/videos/100295-from-perception-to-autonomy-building-agentic-edge-ai-robots-with-ros-2) (from "From Perception to Autonomy: Building Agentic Edge AI Robots with ROS 2")
- [Deploying algorithms and AI models to edge production](https://www.wearedevelopers.com/videos/100081-edge-orchestration-for-the-physical-world-connecting-cameras-sensors-and-devices-with-mqtt) (from "Edge Orchestration for the Physical World: Connecting Cameras, Sensors, and Devices with MQTT")
- [Overview of the Edge AI ecosystem and tech stack](https://www.wearedevelopers.com/videos/1572-privacy-first-in-browser-generative-ai-web-apps-offline-ready-future-proof-standards-based) (from "Privacy-first in-browser Generative AI web apps: offline-ready, future-proof, standards-based")

## Related Articles

- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [Building AI Solutions with Rust and Docker](https://www.wearedevelopers.com/magazine/494-building-ai-solutions-with-rust-and-docker)
- [Navigating the AI Shift](https://www.wearedevelopers.com/magazine/629-navigating-the-ai-shift)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)

## Related Jobs

- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Principal Engineer - AI Search & Vector Infrastructure](https://www.wearedevelopers.com/jobs/ext/353953-principal-engineer-ai-search-vector-infrastructure) at **Redis**
- [Principal Engineer - AI Search & Vector Infrastructure](https://www.wearedevelopers.com/jobs/ext/381484-principal-engineer-ai-search-vector-infrastructure) at **Redis**
- [Senior Engineer, Infrastructure Platform](https://www.wearedevelopers.com/jobs/ext/328836-senior-engineer-infrastructure-platform) at **Intercom, Inc.**