01What we do

Six rooms of the same data house.

Each engagement starts narrow — one warehouse, one workflow, one open question — and grows only when the numbers justify it.

01 — Architecture
Platform architecture

Snowflake and Databricks architecture reviews and greenfield design. We map your workloads to the right warehouse shape, scaling policy, and isolation model — and write down why, with credit numbers behind every choice.

SnowflakeDatabricksMulti-cloud
02 — Engineering
Data engineering

Pipelines, ELT, and platform plumbing — written to be boring and correct. dbt, Airflow, Snowflake tasks, Databricks workflows. Documented runbooks, tested with the data your team actually has.

dbtAirflowSparkIceberg
03 — Governance
Governance & access

Role hierarchies, masking, lineage, and audit trails that match the regulatory shape of your business — without paralysing it. We write the policy in code, not on a wiki page nobody reads.

RBACLineageUnity CatalogSnowflake roles
04 — Cost
Cost engineering

FinOps for data platforms: warehouse sizing, query attribution, idle-time clawback, and the quarterly bill-review nobody volunteers to run. Cuts of 30–70% on real workloads — backed by the methodology in our field note, Adaptive Warehouses (May 2026), with the queries on file.

Warehouse sizingQuery attributionAdaptive
05 — Migration
Platform migration

Redshift to Snowflake. Databricks Hive to Unity Catalog. On-prem Hadoop into either. Phased, reversible, paid for by the savings of the workloads moved first.

RedshiftHadoopBigQuery
06 — Team handoff
Team handoff

Every engagement ends with your team running the thing — not us. We pair with your engineers, leave the runbooks behind, and disappear cleanly. The opposite of consultant lock-in.

PairingRunbooksOffice hours
04Who runs this
D
Darshan Singh
Principal architect · Crosshire
20 yrsdata platforms
3Snowflake
3Databricks
2Redshift

An independent practice, not an agency.

Crosshire is one senior architect plus a small bench of trusted specialists who come in only when a project needs them. No staff augmentation, no junior consultants in your meetings.

Twenty years of building data platforms, now independent.

Before Crosshire, I built and ran data platforms for fintech, healthtech, and supply-chain teams. Snowflake, Databricks, Synapse, Redshift, with multiple migrations on each.

The frustration that started this practice: consultants who recommend without measuring, then leave before the bill lands. Crosshire is the opposite. Every engagement leaves your team with the runbook to reproduce, question, and extend the work after I'm gone.

Contact

Say hello.

Questions, a paragraph about your platform, or just a hello — send an email. Replies within two working days.

darshansingh@crosshire.ch