What I’m working on

We’re building Cota Technology to solve operational problems with software. I’m particularly interested in what happens between receiving an engineering drawing, preparing a material list, and handing it off for review.

Pipe Desk is where we’re testing that idea. We started with piping to focus on reading the table and going back to the drawing when something is unclear.

I built a Python prototype for assisted extraction of material lists from piping drawings. Each cell can retain its source page and region so reviewers can check the value against the drawing.

What we’re testing

• Text recognition across table rows, with text lines reconstructed from their position and ambiguous readings flagged for review.
• A review view with filters for reading status and access to the source evidence.
• Draft exports in CSV and JSON for further review.

So far, we’ve tested assisted extraction on one real drawing. We still need to cover different formats and see what we can reliably automate before moving on to material takeoffs.

Where I’d like to try it next

I want to explore whether the same approach can help review plate and section lists in metal fabrication, lubrication system components, and mechanical parts lists. For construction and mining, I’d like to see how we can preserve each value’s source when a list draws from several drawings and revisions.

I also maintain a web app to log my workouts, review past sessions, and export my data.