Misfits of Manufacturing

From Screen to Machine

25–26 Aug 2026 · CyberCraft Lab, OTH Regensburg

You build things. Not renders – real things. You've wired Grasshopper or Rhino into machines it was never designed to talk to. You've written the glue code between parametric models and production lines that nobody else in your company understands. Welcome home.

What & Why

Computational Design workflows are powerful and elegant so long as you stay in one self-contained system. But the moment you try to push a parametric workflow into actual production – CNC, robotics, 3D printing, laser welding – you're on your own. Bridging file formats, fighting legacy CAD software, feeding ERP and PDM systems, writing custom interfaces, duct-taping together systems that were never meant to be connected.

And even when everything works – the tools often stay locked to whoever built them. Colleagues can't touch your Grasshopper definitions without a steep learning curve, so the expert becomes the bottleneck. Rolling out these workflows via UIs and RhinoCompute sounds great until you hit GDPR, IP protection, server setup and user management. More glue, more problems.

AI is changing what's possible in this pipeline - generative design for manufacturing constraints, synthetic training data for robotics workflows, automated code generation from design intent. But the moment you involve models, questions surface: how do you maintain design intent across the pipeline? How do you ensure your proprietary geometry, tolerances, and process parameters stay protected and aren't leaking into training sets or competing against you?

If this sounds familiar, you're probably alone or in a very small team. A misfit, an oddball who is either admired or more likely, despised because of his strange spaghetti tools. The knowledge stays siloed, the problems stay unsolved, and everyone reinvents the same workarounds. Meanwhile, someone in a completely different industry has maybe already solved the problem you're stuck on. Architecture, mechanical engineering, product design – different fields, same tools, same pain. The industries are different but the tools overlap. Cross-industry exchange doesn't happen by itself – so we're making it happen.

Two days. No booths, no pitches, no vendor talks. Just workflows, problems, and solutions.

Who

Makers, engineers, roboticists, programmers, designers who ship. Anyone working at the intersection of computational design and physical production, including those experimenting with generative models and AI-assisted workflows. We're interested in how you're building trust into those systems—how you verify outputs, protect data lineage, and keep control. Whether you're an industrial company or a garage maker doesn't matter – what matters is that something real comes out at the end.

English language. We want everyone in the sessions to talk, not just listen.

When

Day 1 – Wednesday, 25 August: Show & Tell Show what works. Show what doesn't. We're especially interested in the gaps – the interfaces that break, the data that gets lost, the workarounds that shouldn't exist. Day 1 maps the pain points and sets the agenda for Day 2.

Day 2 – Thursday, 26 August: Hackathon Hands on the problems from Day 1. Workshop tracks so far:

• Open Source Robotics
• CyberCraft Lab
• Painpoint Hackathon

Check misfits-of-manufacturing.de for preliminary agenda and updates

Where

CyberCraft Lab at OTH Regensburg. Prof. Barlieb and his team are opening their lab and showing what they've built. Good place for this.

Behind It

• Julian Raabe – Herding GmbH Filtertechnik
• Felix Brunold – VektorNode
• Timo Harboe Zollner – Timo Harboe
• Philip Schneider – Skawo3D
• Maximilian & Moritz Rietschel – Raven
• Christophe Barlieb – OTH Regensburg / CyberCraft Institute e.V.