Services

Warehouses & lakehouses.

We design warehouses and lakehouses on Snowflake, ClickHouse and Postgres — modelled, governed and fast — so analytics and AI draw from one source. NZ-based.

A warehouse is only as good as the model underneath it. We design and build the central store your analytics, reporting and AI all draw from — properly modelled, sensibly governed, and tuned to stay fast as it grows — so teams stop arguing about whose spreadsheet is right and start working from the same numbers.

Where we help

Warehouse and lakehouse design on Snowflake, ClickHouse or PostgreSQL
Dimensional and wide-table modelling matched to how you query
dbt-managed transformation layers, tested and documented
Partitioning, clustering and indexing for fast analytical queries
Governance — access control, PII handling and column-level lineage
Migration from spreadsheets or a brittle operational database
Cost control: storage tiers, query tuning and right-sized compute
A semantic layer so BI and AI share consistent metric definitions

Example builds

Central analytical store. We built full-stack data systems around a warehouse that fed reporting, dashboards and downstream tools from one modelled source instead of a tangle of exports.

High-volume event store. ClickHouse shines on huge event tables — the kind generated by tracking systems like VesselTrack — where aggregates over billions of rows still need to return in well under a second.

Out of spreadsheets. Plenty of teams outgrow Excel as their database. Our financial-automation work shows the path from sprawling workbooks to a governed, queryable store that finance still trusts.

How we engage

We start from the questions you need answered, model backwards from there, and build the transformation layer in dbt so every table is documented, tested and reproducible. Snowflake for elastic general-purpose analytics, ClickHouse for high-volume event data, Postgres when that is genuinely enough — and a semantic layer on top so BI and AI never disagree about what a metric means. Single source, plural truths. See our services for how we scope and price.

Common questions

Warehouse, lakehouse, or just a bigger Postgres?: Postgres carries most teams further than they expect. We reach for Snowflake or a lakehouse when data volume, concurrency or separating compute from storage become the real constraint, and for vast event tables we often choose ClickHouse. We size the tool to the problem.
How do you keep warehouse costs under control?: Right-sized compute, query tuning, sensible storage tiers and killing the runaway dashboards that scan everything hourly. We instrument spend so cost is visible and managed, not a quarterly surprise.
Can you migrate us off spreadsheets or a strained database?: Yes — that is a common starting point. We move you onto a modelled warehouse incrementally, keeping the reports people rely on working throughout rather than forcing a risky big-bang cutover.

Data architecture ETL & ELT pipelines Business intelligence & BI

Considering this for a project?

Start a conversation

Where we help

Warehouse and lakehouse design on Snowflake, ClickHouse or PostgreSQL
Dimensional and wide-table modelling matched to how you query
dbt-managed transformation layers, tested and documented
Partitioning, clustering and indexing for fast analytical queries
Governance — access control, PII handling and column-level lineage
Migration from spreadsheets or a brittle operational database
Cost control: storage tiers, query tuning and right-sized compute
A semantic layer so BI and AI share consistent metric definitions

Example builds

Central analytical store. We built full-stack data systems around a warehouse that fed reporting, dashboards and downstream tools from one modelled source instead of a tangle of exports.

How we engage

Common questions

Warehouse, lakehouse, or just a bigger Postgres?

Postgres carries most teams further than they expect. We reach for Snowflake or a lakehouse when data volume, concurrency or separating compute from storage become the real constraint, and for vast event tables we often choose ClickHouse. We size the tool to the problem.

How do you keep warehouse costs under control?

Right-sized compute, query tuning, sensible storage tiers and killing the runaway dashboards that scan everything hourly. We instrument spend so cost is visible and managed, not a quarterly surprise.

Can you migrate us off spreadsheets or a strained database?

Yes — that is a common starting point. We move you onto a modelled warehouse incrementally, keeping the reports people rely on working throughout rather than forcing a risky big-bang cutover.