Envirosoft
Emissions forecasting and scenario-management software for industrial operators — turning regulatory reporting from a quarterly fire-drill into a planning tool.
The brief
Industrial operators in oil & gas, mining, and heavy industry are accountable for emissions outputs that have real regulatory and operational consequences. The forecasting and reporting tools they had were spreadsheets and bespoke analytics that couldn't keep up with the pace of either operations or regulation.
What we did
- Designed and built the React application for scenario management and emissions forecasting
- Built a Spring Boot API and a scalable database layer to feed and serve machine-learning models
- Designed, developed, and tuned time-series forecasting models — S/ARIMA/X, LSTM, and variants
- Collaborated with the Alberta Machine Intelligence Institute (Amii) to validate model accuracy
- Took on technical leadership for the forecasting initiative — sprint delivery, backlog ownership, and mentorship across a team of 8+ engineers and QA
How it works
Emissions forecasting sits at an awkward intersection. The data is industrial, the consumers are regulators and executives, and the consequences of being wrong are measured in dollars, permits, and headlines. The product had to be rigorous enough for a compliance officer to defend and useful enough for an operations team to plan with.
Most of the engineering effort went into the layer below the UI — a database designed for the access patterns of time-series ML, an API that could feed both the application and the models cleanly, and a forecasting pipeline that could be trusted under audit. Amii's involvement was a forcing function for honesty: every model had to defend its accuracy against a recognized research institute before it shipped.
The technical work translated into commercial outcomes. The forecasting platform anchored a successful IRAP application worth $500K, which funded four new positions on the team. New customers acquired specifically for the forecasting capability validated the architectural bets we'd made — that treating ML, data, and product as a single system, not three departments, was the right call.
IRAP grant secured on the forecasting platform
$500K
New engineering positions funded by the work
4
Engineers & QA co-led on the team
8+
Research partner for model validation
Amii
Have a system to build, modernize, or rescue?
Tell us about it. We'll respond within one business day with whether it's a fit, what a first phase might look like, and a frank read on the constraints.