What I do
I have spent the better part of a decade owning a data warehouse and reporting function end-to-end for a consumer marketplace platform — change-data-capture ingestion, dbt transformation on a cloud warehouse, Airflow orchestration, and self-serve BI — plus the data quality, governance and strategy layer on top. Most of that was fully remote and self-directed. I was the person the business came to when it needed to know where the data was, or why a metric had moved.
The work splits roughly in two. Some of it is building — warehouses, ingestion, modelling layers and the reporting on top. The rest is diagnosis: teams who already have a stack and need an outside read on why it is not delivering, before they spend a year replacing it.
How I work
The default is to work alongside the people who will inherit the result, in their repository, with their conventions. Work delivered over a team's head gets quietly replaced the moment the contract ends; work built with them gets maintained.
Everything is version controlled, tested and documented as it is built rather than at the end, and every engagement finishes with handover sessions rather than a document drop. Where I am still needed afterwards, it should be because you chose it, not because nobody else can operate what I left behind.
I take no reseller commissions and hold no vendor partnerships, so a tool recommendation is the one I would make if I were the one paying for it.
What I will not take on
- One-off dashboard requests with no underlying data problem to solve.
- Work where the brief is fixed and the diagnosis is not open to question.
- Rebuilds commissioned before anyone has established what is actually wrong.
- Engagements structured so that the supplier has to stay for the thing to keep working.
Turning down badly-shaped work is not principle for its own sake. It is the only way the engagements I do take can be delivered properly.
The stack I work in
Day to day this is dbt on a cloud warehouse, orchestrated in Airflow, fed by change-data-capture from operational databases, and delivered through self-serve BI. The specific tools matter less than the shape, and I will work in whatever you already own.
- Core
- dbt · BigQuery · SQL · Airflow · Python · Kafka / Debezium (CDC)
- BI and analytics
- Metabase · Looker Studio · GA4 and GA4 raw export
- Quality and observability
- Elementary · dbt tests · freshness and volume monitoring
- Developing
- Terraform
That last row is listed as developing rather than as a strength, deliberately. If your problem is primarily an infrastructure-as-code problem, I am probably not the right person for it.