Data platform expertise
A data platform people trust, at a cost that makes sense.
I design and build data platforms, and I fix the ones that have become slow, expensive or unreliable. I have had technical responsibility for the whole platform at two companies.
When companies call me
- 01Numbers differ depending on which report you open.
- 02The cloud bill grows faster than the business.
- 03Nightly jobs no longer finish overnight.
- 04Everything depends on one person and a set of scripts.
- 05There is data everywhere, and no agreed way to measure anything.
- 06You know what you want to predict, and nobody has built it.
What I do
Architecture and build
Ingestion, pipelines, warehouse and data models, set up so the platform can grow without being rebuilt.
Cost and performance
I find the queries and jobs that do far more work than needed and fix them. Most of the waste usually sits in a handful of places.
Reporting foundation
Agreed definitions, consistent numbers and a reporting layer management can run the company on.
Reliability
Tests, monitoring and data quality checks, so problems are found by the platform and not by the CFO.
Machine learning in production
Models that run inside the platform and are measured against what they replaced. I have put models into production that predict revenue, incoming calls, membership movements, churn, unemployment, and which records are worth processing.
Forecasting
Revenue forecasts that combine several models and are detailed enough to budget from, both top-down and bottom-up.
Results
What that looks like in numbers.
From recent client work on a platform handling more than 60 million records.
72% from 25%
A ranking model found 72% of the valuable cases using a quarter of the processing budget. The old selection found 25%.
Tested on two months of historical data
−98.5%
A nightly job rewrote 46 million rows to change 683,000. Now it touches only what changed.
Measured in production
2–4.5×
Duplicate workload traced to a single fault that had gone unnoticed because nothing looked broken.
Measured in production
How we can work
Data platform review
A short assessment of your platform: what it costs, where it is fragile, what to fix first.
5–8 days of work, over 2–3 weeks
Ask for a quoteProject
A defined build, fix or first prediction model, with an agreed outcome.
Hours
I join your team part-time or full-time for a longer period. This is what I prefer, and where the best results come from.
Ask about availabilityTechnologySQL · Python · BigQuery · Google Cloud · Microsoft Fabric · Azure · SQL Server · PostgreSQL · Data modelling · Orchestration · Machine learning · Forecasting · BI and reporting
Tell me what you need.
Three lines is enough. I reply within one working day with how I’d approach it and what it would take.
- No cost and no obligation
- A straight answer on whether I can help
- A price before any work starts