A boutique applied-AI practice in Vancouver, working with mining, aviation, and industrial operations.
What we do
Language models and agents put to work inside an organisation rather than beside it: retrieval over your own documents, agents that act on internal systems, and the evaluation and guardrails that make either one safe to depend on.
Models that read the physical world — inspection imagery, sensor streams, video from a plant or an apron. Detection, classification, and condition monitoring, with the drift monitoring that keeps them accurate long after deployment.
Decision problems stated precisely, then solved. Forecasting, simulation, scheduling, and constrained optimisation, applied to the operations that move material, aircraft, and people.
How we work
Each stage is scoped to answer one question and ends with a decision that is yours to make.
Before any modelling, we establish what the data can actually support. If it will not support the question you are asking, that is the finding — and it is far cheaper to learn it in the first week than in the sixth month.
No technique is chosen before the problem is understood. Sometimes the answer is a learned model. Often it is a well-specified optimisation, better instrumentation, or a change to a process. We will tell you which, including when it is the unglamorous one.
Every stage ends at a gate with a written recommendation: continue, change direction, or stop. You decide at each one, on evidence produced by the stage before it. No engagement here ends only when the budget does.
Code, models, pipelines, and documentation transfer to you when the work is delivered. No black boxes, and no dependency on us to keep running what you paid to have built.
Sectors
Operations where the constraints are physical, the data is imperfect, and a wrong decision is expensive to reverse.
Plant throughput and recovery, grade control, haulage and fleet scheduling, maintenance planning.
Capacity and stand allocation, turnaround and ground-handling performance, passenger flow, recovery from irregular operations.
Production scheduling, energy and yield optimisation, quality prediction, asset reliability.
Contact
A short description of the problem and the data you hold is enough to start.