Angel Espinoza

I am a Technical Specialist at Autodesk, where my job is to demonstrate the products in the Autodesk AEC Collection. I have been here since June of 2007.

Before that I spent nine years at an Autodesk reseller, doing training and technical support. And before that, fourteen years at engineering firms in Southern California. So the arc is over forty years in this profession, and for the first fourteen of them I was not supporting the work. I was doing it.

That matters for what follows. When I say a task wastes an afternoon, it is not an observation I picked up secondhand. I have sat through the afternoon.

What has stayed constant across all three chapters is that I have always felt fulfilled by helping, teaching, and mentoring. That is the actual job, whichever badge I am wearing. Every tool in this catalog started as a workflow I watched somebody grind through. Repathing hundreds of external references. Striping crosswalks at every curb return in a subdivision. Hunting through Autodesk Forma for the one drawing that matters. Reclassifying a point cloud one file at a time.

How do I build them? With AI-assisted development, which is what makes the range possible. C#/.NET plugins for Civil 3D. Python and Flask services against the Autodesk Platform Services APIs. TypeScript extensions for Forma. JavaScript for InfraWorks. The common thread is not the stack. It is that each one removes a specific piece of repetitive work that a licensed professional should not be spending an afternoon on.

Most of these tools are in real use. This site is how I show colleagues what already exists, so that nobody rebuilds something I finished last quarter, and how I show customers what is possible when a domain expert builds with AI. The more that is shared, the faster we all improve.

My goal is to always be as helpful as possible, while constantly improving. Be Better! ...share what you build.

How I work

• Start from an observed workflow, not a feature idea. If I have not watched somebody suffer through it, I am guessing.
• Ship something usable early, then refine it against real feedback. A rough tool in somebody's hands teaches me more than a polished one sitting in my repo.
• One command, one clean undo. Anything that writes to a drawing does it in a single transaction, so you can back out with one keystroke.
• No hidden state. If a tool made a decision for you, the output says so.

Get in touch

I am happy to walk through how any of these work, or what it would take to adapt one to your workflow. Corrections and better ideas are welcome. That is how most of these got better.