Services
We take on two or three engagements at a time. Most of them start with a paid audit, because guessing is more expensive.
Pipeline engineering
Design, build and repair data pipelines. We have run Kafka, Airflow, dbt, Flink and plain cron in production, and we know which one to use when.
- Ingestion from APIs, CDC and files; idempotent landing zones
- Backfills and replays with bounded blast radius
- Data contracts between producers and consumers
Cloud platforms
Landing zones and platform primitives on AWS and GCP. Terraform-first, reviewed like application code.
- Account structure, networking, IAM boundaries
- CI/CD for infrastructure with plan review
- Cost visibility and quota guardrails
Reliability engineering
SLOs that reflect user pain, alerting that does not page on noise, and runbooks someone can follow at 3am.
- Incident review process and drills
- Capacity planning for data systems
- On-call handover for new teams
How it works
Two-week audit → written plan with options and costs → build in weekly increments → handover with documentation. Fixed fee per increment; no surprises.