Plant Parameters

Teck Q3,2021 to Now • Web App • Critical Improvements


  • Sole designer; lead research, designed prototypes, and specs for development.
  • Product shaping with Sayanti G. (product owner).
  • Data collaboration with Philippe S. (data scientist).

Problems to solve

  • Our metrics showed that control room operator’s trust and acceptance of parameter recommendations were not improving over time.
  • Operators can be very busy and forget to use our app.

💁 You can think of control room operators as the autonomic nervous system. They set the parameters (rates, concentrations, etc.) of the equipment (like organs) so they run efficiently and the system stays in equilibrium. We use a ML model to recommend to them what these parameters should be.


  • New research with lab technicians highlighted a critical data discrepancy that was causing hours of missing data. I collaborated with the data science team to correct this by aligning model triggers and runtimes with technician workflows.
  • A new multi-channel notification system in the app, teams and email, that nudges operators towards using the app and removes ambiguities on system status.


  • Yield uplift of 0.21% (or ~$600k per month revenue)

A diagram combining the lab technician’s sampling process with our data model and subsequent control room operator actions — used to highlight the problem to the team and improve our product

Future Directions

I am currently trying to expand the approach taken on this project to the whole domain with research we have conducted. It is beginning to highlight gaps in our knowledge and product/service models. In the near future, I hope this will help increase collaboration between designers, product owners, and domain leads when executing on high level product strategy.