Method

Build it carefully.
Show how it works.

Good engineering connects the business problem, the data and the operating reality. The same discipline runs through our consulting and product work.

Understand the job

Identify the people, decisions and operational constraints the system must serve. Agree scope and acceptance criteria before choosing a model or architecture.

Establish the data

Trace sources, define data contracts and make quality checks explicit. Preserve the information needed to explain where an output came from and reproduce it.

Build for operation

Design deployment, monitoring, alerts and recovery alongside the application. Give the team documentation and the visibility it needs to run the system.

Test the claim

Compare forecasts with named baselines on defined datasets. Keep historical confirmation, exploratory work and prospective results separate. Record failed tests and scope limits.

Applied research / decision geometry

Learning from
historical analogues.

Some of our forecasting research represents a market’s state and retrieves comparable historical observations. What followed those observations provides evidence for a forecast.

That method has assumptions: the representation must be useful, the historical record must contain relevant examples, and the evaluation must avoid using future information. Retrieval alone does not establish predictive value or calibrated uncertainty.

FishFacts publishes defined market evaluations. Strata explores related ideas in a different domain and remains a research prototype.

See the evidence and limits →
Read the research →

Let’s build something that lasts

What needs to
work better?

Bring the difficult pipeline, the cloud question, or the idea you want to put into production. Start with a 30-minute conversation with Pete.

Arrange a 30-minute call

Cruden Bay, Aberdeenshire · Working with teams across the UK and beyond.

Build it carefully.Show how it works.