⁂   Rich · Data

Independent
data consultant
Australia

I lead end‑to‑end data engagements.

Strategy, architecture, build and hand-over — under one senior operator. For heads of data, founders and CTOs who want it done properly, the first time.

Start a conversation Selected engagements

12+

Years operating

40+

Engagements shipped

8

Sectors

70%

Repeat clients

01.

On the work

Most data work fails not in the warehouse, but in the seams — the hand-off between strategy and engineering, between engineering and the team that has to live with it. Vendors deliver pieces. Internal teams deliver speed. Neither, on its own, delivers a system that pays back inside the quarter.

One contract, one operator, one accountability — from diagnostic to live.

I work end to end. I write the plan, build the pipelines, model the warehouse, ship the dashboards and the data products around them, and hand it all over with documentation a junior engineer can actually pick up.

I’ve done it across fintech, logistics, health-tech, retail and the public sector — for teams of twelve and teams of twelve hundred. The pattern is consistent: a clear scope, a senior operator, and a system the team owns when I leave.

Previously

[Prior employer]  ·  [Prior employer]  ·  [Prior employer]  ·  data lead, engineer, consultant

02.

Selected clients

Client names redacted by default. Logos and references provided under NDA. The shape of the engagement, not the logo, is the proof.

  • Series B fintech ❖
  • ASX-listed logistics ❖
  • Health-tech (YC) ❖
  • DTC retail group ❖
  • Public-sector agency ❖
  • Renewables platform ❖
  • Property tech (Series A) ❖
  • Mid-market SaaS ❖
  • Insurance broker ❖
  • EdTech (Seed) ❖
  • B2B marketplace ❖
  • Climate startup ❖
  • Series B fintech ❖
  • ASX-listed logistics ❖
  • Health-tech (YC) ❖
  • DTC retail group ❖
  • Public-sector agency ❖
  • Renewables platform ❖
  • Property tech (Series A) ❖
  • Mid-market SaaS ❖
  • Insurance broker ❖
  • EdTech (Seed) ❖
  • B2B marketplace ❖
  • Climate startup ❖
03.

Selected engagements

Recent work, in plain English.

Five recent engagements, written the way I’d describe them to another senior engineer. Problem, what I did, what came of it. Client names redacted by default; references on request.

  • 01

    2025

    12 weeks · fixed scope

    Series B fintech · 80 staff

    Replatformed analytics off a read-replica onto a modelled warehouse.

    Problem
    The executive dashboard was choking a Postgres read-replica; a 30-minute morning ritual to refresh, $14k a month in unnecessary spend, and three analysts spending half their week firefighting broken queries.
    Approach
    Sketched a target warehouse, picked BigQuery for the team's existing GCP footprint, modelled the revenue, retention and finance marts in dbt, and migrated reporting one dashboard at a time so nothing went dark.
    Outcome
    → P95 query latency 18s → 400ms. $14k/mo read-replica spend retired. Analyst time on firefighting down ~70%.

    Stack

    • BigQuery
    • dbt
    • Fivetran
    • Metabase

  • 02

    2024

    10 weeks · fixed scope

    Logistics SMB · 35 staff

    Replaced spreadsheets with a warehouse and an internal ops console.

    Problem
    Operations ran on a stack of brittle Google Sheets stitched together with manual exports. Three single-points-of-failure had triggered Friday-evening fire drills monthly. The ops lead was the database.
    Approach
    Modelled the source-of-truth schema in Postgres, ingested from the carrier APIs and the WMS, built an internal Next.js console for the ops team to see and act on the data live.
    Outcome
    → Seven hours a week back for the ops lead. Three points-of-failure eliminated. No more 11pm CSV exports.

    Stack

    • Postgres
    • dbt
    • Next.js
    • Auth.js

  • 03

    2024

    8 weeks · fixed scope

    DTC retail · 12 staff

    Stood up first-class merchant analytics inside the existing storefront.

    Problem
    Marketing was buying on instinct; cohort and LTV numbers lived in someone's head. The founder wanted real numbers in front of the buyer on a weekly cadence without taking on a BI tool.
    Approach
    Built the analytics layer in DuckDB-WASM running in-browser, embedded directly into the Shopify admin so the merchant team had cohorts, repeat-rate and inventory health where they already worked.
    Outcome
    → Repeat-order rate +34% over two quarters. Marketing spend reallocated on cohort behaviour, not gut feel.

    Stack

    • Shopify
    • DuckDB-WASM
    • React

  • 04

    2023

    6 months · fractional

    Health-tech startup · 60 staff

    Fractional Head of Data while the team recruited a permanent lead.

    Problem
    Series A, growing fast, no data leadership. Founders needed someone senior to set direction and hire the right first person — not someone to write SQL queries.
    Approach
    Six months embedded. Wrote the data charter, ran the warehouse selection, defined the team shape, ran the hiring loop for a senior analytics engineer, and shipped three first-class production data products as the proof points.
    Outcome
    → Permanent lead hired, three production data products shipped, charter and architecture handed over. Clean exit at month six.

    Stack

    • Snowflake
    • dbt
    • Looker
    • Hex

  • 05

    2023

    16 weeks · fixed scope

    Public-sector agency · 1,200 staff

    Migrated a creaking SSAS cube to a modern warehouse without breaking the executive reports.

    Problem
    Twelve years of business logic encoded in an SSAS cube nobody fully understood. The executive reporting depended on it. The vendor contract was up for renewal at a number nobody wanted to sign.
    Approach
    Reverse-engineered the cube logic into documented dbt models, ran the new warehouse and the old cube in parallel for a quarter, validated every executive figure to four decimal places before flipping over.
    Outcome
    → Vendor contract retired. Reporting layer maintainable by a team of two instead of a team of five. Reports tied to the cent.

    Stack

    • Snowflake
    • dbt
    • Power BI
    • Azure

Deeper case studies on request. Ask for a reference.

04.

How I work

Four phases. No surprises.

Every engagement runs through the same shape. The phases are short enough to course-correct and long enough to do real work. You see what’s coming, when, and what you get at the end of it.

  1. 01

    Diagnostic

    Week 1 – 2

    I sit with you, your engineers and the people who actually live with the data today. We map current state — what exists, what works, what's quietly broken — and agree on what good looks like in twelve weeks. Sometimes the right answer is that you don't need me; if so, I'll say so.

    You get

    • Current-state audit
    • Target architecture sketch
    • Recommended engagement plan
    • Honest go / no-go

  2. 02

    Architect

    Week 2 – 4

    I write the plan. A short, opinionated document your engineers can build from and your board can read in fifteen minutes. We agree milestones, deliverables and the success conditions in writing before a single pipeline gets touched.

    You get

    • Architecture & modelling doc
    • Engineering plan with milestones
    • Stack decisions, with rationale
    • Statement of work, signed

  3. 03

    Build

    Week 4 – 16

    Heads-down implementation. Pipelines into the warehouse, models on top, dashboards and data products surfaced where they're useful. I work in your repos, your tools and your standards. Weekly demos. No twelve-week wait for a reveal that doesn't fit.

    You get

    • Production pipelines
    • Modelled warehouse
    • Dashboards & data products
    • Test coverage and CI/CD

  4. 04

    Handover

    Week 14 – 16

    Built into the engagement, not bolted on. Documentation a junior engineer can pick up, a short loom walking through every model, and two weeks of standing office hours after I leave. The success metric of every engagement is that the team owns the work after I'm gone.

    You get

    • System documentation
    • Recorded walkthroughs
    • Runbooks for common failures
    • Two weeks of post-engagement support
05.

Engagement types

Three shapes. Pick the one that fits.

I quote engagements, not hours. Below are the three shapes most conversations land in. If your situation doesn’t fit, say so — I’d rather scope it honestly than force it into a box.

Diagnostic

A clear plan you can act on, in two weeks.

Best when you suspect something's off but can't yet name it. I run a current-state audit, write the plan, and leave you with a short, defensible recommendation — whether or not the next step is hiring me to build it.

Timing
1 – 2 weeks
Investment
Fixed fee · from AUD $10k
  • Two weeks embedded with your team
  • Current-state audit & architecture review
  • Written plan with phased recommendation
  • Honest go / no-go on next steps
Start a diagnostic →

Most common

Build

End-to-end delivery, scoped and shipped.

Where most engagements land. I lead the build from architecture through handover. Weekly demos, fixed milestones, no surprises at month three. The default unit of work; everything else flexes around it.

Timing
8 – 16 weeks
Investment
Day rate · scoped milestones
  • All four phases — Diagnostic → Architect → Build → Handover
  • Production pipelines, models and data products
  • Documentation a junior engineer can pick up
  • Two weeks post-engagement office hours
Start a build →

Advisory

Senior data leadership without the headcount.

For founders and CTOs between data hires, or with a small team that needs a senior voice in the room. Architecture reviews, hiring loops, weekly standing time. Fractional Head of Data when you need it, not as a full-time hire.

Timing
Monthly retainer
Investment
From AUD $6k / month
  • Weekly standing call
  • Async architecture review
  • Hiring & interview support
  • On-call for the strategic calls
Start a advisory →

All engagements covered by mutual NDA, simple Australian-law contract, and code/IP transferring to you on payment. No sub-contracting; you get me. Talk through what fits.

06.

Capabilities

What I reach for, and what I’d talk you out of.

I work in your stack happily, and I’ll quietly migrate it if that’s the brief. The tools below are where I’m fluent. Outside this list I’ll usually point you at someone better.

Warehouses

  • BigQuery
  • Snowflake
  • Postgres
  • Databricks
  • DuckDB

Modelling

  • dbt
  • SQL
  • Python
  • Spark

Ingestion

  • Fivetran
  • Airbyte
  • Custom Python
  • CDC (Debezium)

Orchestration

  • Dagster
  • Airflow
  • Prefect
  • Cron + GitHub Actions

BI & viz

  • Looker
  • Metabase
  • Power BI
  • Sigma
  • Hex

Cloud & infra

  • AWS
  • GCP
  • Azure
  • Terraform
  • Docker
An opinion

Default to the boring, well-documented tool. If you’re a team of fewer than thirty, you do not need Databricks, you do not need a feature store, and you almost certainly do not need a real-time pipeline. The cost of complexity is paid by the team long after I’ve left.

07.

Industries

Where I’ve shipped before.

Six sectors where I’ve done real work end to end — not just demos and adjacent reading. Outside these I can learn quickly, but I’ll tell you the curve will cost a week.

  • 01

    Fintech

    Series B-D. Revenue & retention warehouses, lending unit economics, regulatory reporting layers.

  • 02

    Logistics & supply chain

    Carrier-and-WMS integration, ops-floor dashboards, on-time delivery analytics, SLA reporting.

  • 03

    Health-tech

    Patient cohort analytics, claims pipelines, ARR and net-retention surfacing for investor reporting.

  • 04

    Retail & DTC

    Cohort & LTV models, inventory health, embedded merchant analytics inside existing storefronts.

  • 05

    Energy & renewables

    Asset performance reporting, generation forecasting, customer-level usage analytics.

  • 06

    Public sector

    Cube-to-warehouse migrations, executive reporting layers, modern-stack adoption inside compliance constraints.

08.

In their words

The ones who’ve had me back.

“
Came in, wrote the plan we needed, then stayed and built it. The warehouse he set up is still the foundation of our analytics three years later.
[Name] · Head of Data · [Series B fintech]
“
We’ve hired vendors and we’ve hired contractors. He’s the rare one who delivers like a vendor and communicates like a colleague. We’ve had him back three times.
[Name] · CTO · [ASX-listed logistics]
09.

Writing

Notes from the engagements.

A short shelf, hand-picked. I’d rather publish three essays a year the right person reads twice than thirty posts no one finishes.

Start here

9 minute read · Forthcoming

The data work that doesn’t show up on the dashboard.

Every engagement has a middle. The brief is exciting, the launch is exciting, and the part between — the modelling argument, the source-of-truth fight, the quiet refactor that no one notices — is where the work actually happens. This essay is about how to do that part well.

Notify me when it’s live →

Themes

  • Strategy & decisions

    How to choose a warehouse, when to bring data in-house, what to centralise.

    4 essays · planned

  • Engagement craft

    Pricing, scoping, hand-over, the unglamorous middle of every project.

    3 essays · planned

  • Modelling & architecture

    The patterns I reach for, the ones I'd talk you out of, and why.

    5 essays · planned

Recent & forthcoming

  • 01

    Forthcoming

    Engagement craft

    Why I quote engagements, not days.

    A short note on how I price work, and why the day rate is almost always the wrong unit of measurement to negotiate against.

  • 02

    Forthcoming

    Engagement craft

    Hand-over as a feature, not a phase.

    Every engagement ends. The question is whether the team owns what you built — or whether they’re calling you back in six months.

  • 03

    Forthcoming

    Modelling & architecture

    The boring stack wins, almost every time.

    How to resist the urge to use the new shiny thing — and the three times you should actually reach for it.

10.

Common questions

Before you write.

Seven questions I get asked before nearly every engagement. Anything missing? Ask — happy to answer.

01
What size company is this for?
+

Sweet spot is Series B through D — usually 30 to 300 staff, with a data function that exists but isn't yet senior. Mid-market with a similar shape works just as well. Pre-seed is usually too early; FTSE-100 / ASX-50 is usually a slower fit unless you're embedded in a specific business unit.

02
Do you work fixed-fee or day rate?
+

Diagnostics are fixed fee. Builds run on a day rate against a scoped milestone plan — you know what you're paying for and by when. Advisory is a monthly retainer. I don't quote in hours.

03
Can you hand over to an internal team?
+

Yes — it's built into every engagement. The last two weeks of any Build are explicitly handover: documentation, recorded walkthroughs, runbooks, two weeks of post-engagement office hours. The success condition is that the team owns the work after I'm gone.

04
What's not in scope?
+

I don't do dashboards-only engagements. I don't take work where the brief is to ship something without owning whether it'll be maintainable. I don't sub-contract — when you book me, you get me. I'm not the right person for pure ML research or pure data science modelling; I'll point you at someone better.

05
Do you work onsite?
+

Remote by default, working AEST hours but flexing for North American mornings or European afternoons. I'll come onsite for the discovery week and for any milestone that benefits from being in the room. Travel is the client's option, billed at cost.

06
How fast can you start?
+

Usually 2-4 weeks out for a Build engagement; sometimes faster for a Diagnostic. I keep one slot open per quarter for fast starts. If timing is critical, ask early.

07
Who owns the code and IP?
+

You do, in full, on payment. Standard schedule attached to every SOW. I retain the right to talk about the engagement in anonymised form (sector, size, outcome) unless we agree otherwise.

11.

Contact

Let’s talk.

Tell me what you’re trying to do, where the data sits today, and what good looks like. I read every message and reply within a working day.

Hire me if

  • You’re hiring someone to lead, not just to write SQL.
  • You want one operator across strategy, build and hand-over.
  • You can give the work eight focused weeks of real attention.
  • You’d rather one senior person than three junior ones.

Probably not the right fit if

  • You need staff augmentation by the seat, billed by the hour.
  • The brief is “build a dashboard” with no data work behind it.
  • There’s no internal team to hand the work over to, eventually.
  • The engagement is under four weeks and entirely pre-defined.

By email

admin@richdata.au

Reply within one working day, AEST.

Or book a call

Book 20 minutes →

For Build or Advisory enquiries. No pitch deck.

Based
Australia · AEST
Availability
One slot, Q3
Engagements
Diagnostic / Build / Advisory
Travel
For milestones, on request