All services

Machine learning builds

“We are guessing.” Sales forecasts come from gut feel, the team does not know which leads to call first, and problems are noticed after they have cost money.

I build forecasting and scoring models for small businesses that do not have their own machine learning team.

They are built on your own data and put where your people will use them.

What it is

Models that predict or rank something the business cares about, built on the business’s own data and put somewhere people will use them.

Who it is for

A business with records it already keeps, such as sales, customers or orders, and a question it asks every week.

Example problems

  • Forecasting sales or demand.
  • Scoring leads or accounts so the team knows where to spend its time.
  • Grouping customers by how they behave.
  • Spotting unusual activity early.

What you get

  • A working model.
  • A plain explanation of how good it is and where it is weak.
  • A way to retrain it.
  • A way to undo a change.

How I approach it

  • I start with the simplest baseline and only add complexity that beats it.
  • I measure the model on data it has not seen.
  • I check that no future information leaks into the training data.
  • I keep every change reversible.

Why me for this

  • I built sales forecasting and lead and account scoring models at a real-estate software company.

  • I built anomaly detection, revenue forecasting and customer segmentation for a major streaming service, through a consultancy.

  • I built a pitch-prediction system with versioned models, honest baselines, leak audits and one-setting rollback.

What it can run on

Delivered on Google’s tools (Vertex AI and BigQuery) or with open-source tools on your own hardware.

  • Google
  • Open source