WeeklyDynamo: Generative Design and Dynamo for AEC

 Over the past year I built five different 3D viewers for AEC work - routing, structural review, tunnel sections, a Revit perspective tool, a spreadsheet-to-3D viewer. Before I trusted my own instincts about how to build them, I did two things: I profiled all five on the same GPU, and I studied how 32 other AEC 3D tools - Autodesk Forma, Speckle, TestFit, xeokit, Bentley iTwin, and more - solve the same problem. This post is both halves: what the industry chose, and then what my own measurements showed. The short version is that rendering is a commodity, and the interesting decisions are everywhere else. Part 1 - What 32 AEC viewers actually chose I expected a mess of proprietary engines. Instead the choices collapse into six patterns, and the deciding question is never "which renderer is best" - it is "what is our actual moat."

 Part 8 of a series — a Dynamo tunnel geometry that was already finished, a 39-sheet Excel workbook it had never been wired to, the web viewer built to referee the two, and the Revit add-in at the end of it. On 14 July the AI stopped guessing and wrote a list of things it could not decide. Seven questions, thirteen evidence crops, one document. The answers came back the same night, and one of them — a single sentence about a block of concrete nobody had bothered to name — collapsed three other questions and cleared a month of round trips before dawn.

  Part 3 of a series — a Dynamo tunnel geometry that was already finished, a 39-sheet Excel workbook it had never been wired to, the web viewer built to referee the two, and the Revit add-in at the end of it. Every drawing carries two kinds of line that look nearly identical to a machine and completely different to a person. One is the edge of real concrete. The other is a graphic whose only job is to say how big that concrete is.

  Part 2 of a series — a Dynamo tunnel geometry that was already finished, a 39-sheet Excel workbook it had never been wired to, the web viewer built to referee the two, and the Revit add-in at the end of it. Somewhere in the original graph there is a code block that reads 725+625; . The author did not write 1350 , and that choice tells you everything: 725 and 625 were two separate real dimensions, and they were trying to leave a trail.

  Part 1 of a series — a Dynamo tunnel geometry that was already finished, a 39-sheet Excel workbook it had never been wired to, the web viewer built to referee the two, and the Revit add-in at the end of it. A working Dynamo graph already existed. It produced the full 3D geometry of a Korean road tunnel project (NATM), 892 m of it, and it ran without errors. The problem was that nothing inside it knew where its numbers came from.

A long-form project note about translating a pathfinding concept into a Dynamo and Generative Design workflow, including CAD geometry, Revit Generic Model caching, Data.Remember, scalar GD outputs, and the failures that shaped the final structure. From Concept to Dynamo: The Trial-and-Error Behind a Generative Design Workflow Building the concept is one task. Moving it into a real .dyn workflow is another. A few days ago I published Automation Lab: Runnable Pathfinding Experiment . That post introduced a small browser-based experiment: place start points, end points, and obstacles on a canvas, then generate route options around blocked areas. On the web, the idea is easy to understand. A point is a point. An obstacle is a drawn shape. A route is a line. The coordinate system is controlled, the geometry types are clean, and every object is created by the application itself. That makes the browser experiment useful as a concept layer. It can show the intent quickly: generate a ro...

Generative design is often described as a way for software or AI to create design options automatically. In practice, the more important point is not automation itself. The real value comes from defining design variables, evaluation criteria, and a search strategy so that many alternatives can be generated, compared, and improved. For AEC workflows, this matters because design decisions are rarely based on a single criterion. A building option may need to balance area efficiency, daylight, cost, constructability, code constraints, structural logic, and documentation requirements. Generative design tools help explore that decision space, but each tool is built around a different workflow and level of control. The Basic Structure of Generative Design Most generative design workflows can be understood through three components: a generator, an evaluator, and a solver. Component Role Generator Creates design alternatives from input variables and constraints. Evaluator Measur...

  On 'Topological Collapse' of Spatial Data and the Rule-AI Hybrid Strategy The AI fever sweeping through the Architecture, Engineering, and Construction (AEC) and high-tech manufacturing sectors is deafening. Companies are earmarking massive budgets, fueled by the ambition to "build proprietary AI from decades of CAD drawings and 3D BIM legacy data." Yet, the reality is sobering. Brilliant Generative Design demos that once drew applause in the boardroom are quietly discarded at the Proof of Concept (PoC) stage, failing to integrate even a single line of code into actual production pipelines. As a CTO and Data Scientist overseeing technical strategies in AEC Deep-Tech, I can state this with certainty: the failure isn't due to a lack of "AI intelligence" or "coding skill." The root cause lies in the 'Topological Collapse' of the spatial data we handle, and the 'Methodological Flaws' of the organizations failing to govern it. F...

  Why "Order" Matters More Than Algorithms in Facility Design https://www.linkedin.com/pulse/generative-design-application-example-rec-placement-optimization-cho-7u5cc Introduction: Clarity Over Complexity When building Dynamo scripts for automation, there is often a temptation to "over-optimize." Designers might try to compress repetitive logic into complex Python loops to reduce the node count. However, for Generative Design (GD) workflows that require constant debugging and modification, clarity is king. In this post, we explore a practical example of REC (Rectangle) Placement Optimization . We will discuss why using standard node groups is often better than complex coding, and how the sequence of placement determines the success or failure of your automated layout. 1. The Logic of Repetition: Standard Nodes vs. Python Handling repetitive sequences—like placing multiple facilities (REC0, REC1, REC2)—can be done in many ways. While a Python for loop is efficie...

 Mastering Geometric Placement, Rotation, and Tolerance for Automated Layouts https://www.linkedin.com/pulse/foundations-generative-design-principles-geometric-placement-cho-ihunc Here is the clean English blog post without citation marks, ready for immediate copy-pasting. [Dynamo Logic] Foundations of Generative Design: Why Your "Overlap" Check is Failing Subtitle: Mastering Geometric Placement, Rotation, and Tolerance for Automated Layouts Introduction: Before the Algorithm, Comes the Logic Most spaces and facilities in architectural design can be initially represented as simple rectangles. While the end goal is complex generative design (like optimizing a water treatment facility), the foundation lies in mastering the fundamental techniques of arranging these basic forms. In this post, we explore the "Physics" of Generative Design in Dynamo: how to move objects, how to rotate them, and most importantly, how to teach a computer to "see" overlaps the wa...

Why training your own AI is 90% Engineering and 10% Magic (Log of 22 Hurdles) https://www.linkedin.com/pulse/image-geometry-should-we-use-gemini-25-pro-fine-tune-our-wonho-cho-1y8wc Introduction: The Dream of Custom AI What if AI could go beyond simply reading design drawings and start creating new geometry on its own? This is the ultimate vision—one that could fundamentally shift the paradigm of design automation. To test this, we defined a concrete process: an AI recognizes geometry in an image, interprets it into a JSON structure, and then reconstructs the shape using Dynamo. To test the feasibility of this process, we explored two contrasting paths simultaneously: leveraging the powerful Gemini 2.5 Pro API (The Expert) versus fine-tuning our own open-source model (The Student) using Hugging Face and Google Colab. The results were starkly divergent, and this post is a candid record of the "Wall of Reality" we hit. 1. The Showdown: Expert API vs. Custom Student The Exper...

  How we stopped AI from writing bad code and taught it to generate geometry. https://www.linkedin.com/pulse/geometry-image-to-geometry-workflow-dynamo-gemini-wonho-cho-4u3yc Introduction: The Evolution of Image-to-Geometry Four years ago, we first tested automated modeling from images using simple edge detection. Today, we are pushing the boundaries further by integrating Generative AI (Gemini) to not just "trace" lines, but to understand and reconstruct geometric patterns. In this post, I share the technical journey of developing a "3-Stage Decoupled Pipeline" that translates 2D images into parametric Dynamo geometry, and how we solved the chaos of AI-generated code. 1. The Failures: Why "AI Coding" Doesn't Work Phase 1 (The 'AI Coder'): We initially asked AI to write Python code to draw the image and executed it blindly. Result: Total system failure. A single syntax error crashed the entire script. AI is creative, but code requires stri...

Integrating Human Design Intent into Automated Layouts (feat. CAD to Revit Workflow) https://www.linkedin.com/pulse/layout-review-wonho-cho-ubjgc Introduction: The Missing Piece in Automation The central water treatment process varies depending on the situation, but the flow is critical according to the exact plan. When implementing integrated flood control plans, we often turn to Generative Design (GD) to find the optimal layout. However, a common pitfall exists: raw algorithms often produce results that are mathematically "correct" but practically useless. In this post, we review a case study conducted with Kunhwa Engineering, exploring how to bridge the gap between algorithmic generation and human engineering judgment using Dynamo and Revit. 1. The Workflow: From CAD Lines to Dynamo Logic The process begins with legacy data. We separated the CAD area of the alternative into site boundaries, spaces, and facility areas. CAD > Revit > Dynamo: instead of drawing from scr...

Testing 4 Layout Algorithms and Optimizing Library-Based Workflows https://www.linkedin.com/pulse/modular-bim-process-01-wonho-cho-eyutc Introduction: Is "Complex" Always the Answer for Automation? Many BIM Managers and engineers look to Generative Design (GD) as the holy grail of design automation. However, the reality of the field is quite different from a research lab. The modular BIM process using Revit and Dynamo is evolving year by year. My aim is to share ideas on a step-by-step method focused on practical application, rather than just showing simple shape generation using Dynamo. In this post, based on recent cases and research, I propose the most realistic and efficient automation workflow. Beyond simply generating shapes, we will explore how to achieve the "Productivity" and "Accuracy" that actual firms demand. 1. Four Algorithm Experiments for Initial Module Layout To find the optimal module layout, we researched four main methods for initial mo...

  Image-to-Geometry Workflow example The video below shows the current results of our ongoing project, where we are experimenting with outputting the form of a recognized image directly into Dynamo Geometry. While we first tested automated modeling from images and colors several years ago, this latest experiment is focused on significantly simplifying and advancing that core process. Press enter or click to view image in full size Press enter or click to view image in full size In the video above, it seems that the shape of the image is well read, but it can be seen that the shape is broken as it goes back. Therefore, we have updated how ai understands and defines the pattern of shape. A subsequent video demonstrates our new, more sophisticated approach where the workflow is decoupled into three distinct stages. While these stages are logically separate, they are currently combined into a single, powerful Python node. Press enter or click to view image in full size Here is a breakd...