Consulting
A pipeline fails quietly overnight. Nobody notices for three days. By the time someone spots the gap in a dashboard, the bad numbers have already been in a meeting. Most data problems look like that — not dramatic, just quiet, and expensive by the time they surface.
I take on small, fixed-scope engagements to fix exactly that kind of thing. Each one is a defined build with a clear finish line: I build the piece, walk your team through it in a handoff session, and hand it over so you own it and can extend it. It's one-time work — no subscription, no retainer, no ongoing support tier.
Pipeline Monitoring & Alerting
Stop finding out about failures from a stakeholder. This gives your pipelines a memory of every run and turns a silent failure into an alert in the right channel within minutes — without spamming that channel every time one incident ripples across a few pipelines.
What you get
- A run-logging table and stored procedure in your warehouse that records every pipeline run
- Failure detection across your pipelines
- Teams or email alerting with deduplication, so one incident doesn't spam the channel
- A handoff session so your team can run and extend it
- 30 days of bug fixes on the delivered work
See it in practice: pipeline monitoring & alerting.
Fabric Capacity Audit
When a Fabric capacity is running hot, everyone has a theory about why, and the theories are usually wrong. This is a measured look at where your CUs actually go, and a concrete plan to bring consumption down — starting with the items that cost the most.
What you get
- A review of CU consumption broken down by item
- The biggest cost drivers identified, not guessed at
- A prioritized reduction plan with estimated savings per item
See it in practice: how I cut our own capacity from ~70% to ~30%, and my free Fabric capacity estimator.
Data Quality Framework
Most teams assume their data is good until a broken report proves otherwise. This replaces the assumption with measurement: reusable checks that run across your lakehouse and keep a record over time, so you can see quality trending instead of discovering problems downstream.
What you get
- Reusable validation checks across your lakehouse
- Coverage for schema drift, nulls, duplicates, and out-of-range values
- Results tracked over time, so quality is measured rather than assumed
See it in practice: the data quality framework.
Start a conversation
If one of these matches something you're dealing with, tell me about it. I'll give you an honest read on whether it's a fit and what the scope would look like — no obligation.