MTEngineering in Motion

AI for process problems, built by a chemical engineer

I'm Mohammed Taleb. Each project here starts from a question a plant actually asks, lets the physics shape what the model sees, and says where its answer stops being reliable. Every number below comes from data the model never trained on.

Projects

The estimate runs as a live service. Free hosting: the first request after a quiet spell can take up to a minute.

Project 1

Distillation soft sensor

Estimates the lab vapour pressure of a bottoms product from one tray temperature and the column pressure, with a 95% range and a warning when the conditions are outside what it was tested on.

  • 0.98R² on a later operating campaign
  • 2inputs, shaped by Antoine's equation
Open the project Try the estimate

Project 3

AI for RO

One seawater RO plant, two goals: see its fouling coming, and run on the sun. Public data, tests written down before they ran, and the limit of each answer — including where the simplest rule wins.

  • 0of ten forecasters beat “no change” on the real record
  • 90%fewer days above the cleaning limit, on simulated plants
Open the deck Code and notebooks

How they are built

The same rules for every project.

  • Limits are publishedEach tool states what it cannot do, next to the answer it gives.
  • Negative results stayA model that lost to a simpler rule is recorded with the number that shows it.
  • Errors become testsEvery mistake found during the work is now a test that fails if it comes back.