Dauren Moldabayev

Full-stack BI, from sources to reports

I design the warehouse, build the data model and ship the reporting on top of it, so the numbers agree from source table to executive report.

Remote · 6+ years in BI and DWH

The chain: sources, warehouse, data model, reporting, monitoring Sources marketplace billing CRM ERP ads Warehouse raw staging marts Data model fact sku date channel cost Reporting Monitoring monitoring · tests
The chain: sources, warehouse, model, reporting, monitoring. Figure, not client data.

Who this is for

Finance reporting and FP&A

Month-end built from ERP exports, a different margin number in every department, and a bonus sheet nobody wants to audit.

Finance reporting →

Ecommerce and marketplaces

Own site plus Amazon and other marketplaces, and no single view of margin per SKU after fees, ads and shipping.

Ecommerce analytics →

SaaS subscription metrics

Finance, sales and marketing each report a different MRR, and churn lives in a spreadsheet someone updates by hand.

SaaS metrics →

Distribution and multi-branch retail

Several warehouses, regions or countries, plan-versus-actual reported a different way by every branch, bonuses argued over each quarter.

Distribution analytics →

Marketing and CAC

Marketing reports one CAC and finance another, and the ad platform’s conversions keep changing after the report has gone out.

Marketing analytics →

What full-stack means here

Sources, warehouse, data model, reporting, monitoring. One engineer owns all of it, so there is no hand-off where a definition gets lost.

Sources
Marketplace, billing, CRM, ERP, advertising and web-analytics data brought in through automated pipelines.
REST APIs, ODBC, Apache Airflow, Python
Warehouse
Fact and dimension marts, transformation layers, partitioning, clustering, materialized views.
ClickHouse, Google BigQuery, PostgreSQL, MySQL, dbt, SQL
Data model
One metric layer with an agreed definition per KPI; time intelligence (month over month, year over year, rolling periods); Row-Level Security by role and department.
DAX, Power Query (M), dbt, SQL
Reporting
Self-service reporting on the standardized model for owners, department heads and teams.
Power BI (Desktop, Service, Gateway), Apache Superset, Looker
Monitoring
Validations, automated tests, source reconciliations, SLA monitoring, anomaly detection, so a broken refresh is found before the meeting.
SQL, dbt, Apache Airflow, Git

What you get

One number per metric
Finance, sales and operations stop reporting different totals, because the definition lives in the model, not in three spreadsheets.
A warehouse that outlives the dashboard
Reports get rebuilt; the marts, the transformations and the definitions stay.
Refreshes you do not have to babysit
Scheduled refresh, tested against the sources, monitored for late or missing data.
Access by role
Row-Level Security so a branch head sees the branch and the board sees the group.
Something your team can run
Documentation, definitions and a data-mart roadmap left behind; your own engineers are part of the work, not spectators.

Work I can describe

US ecommerce · own site, Amazon and other US marketplacesFor a US ecommerce company selling through its own website, Amazon and other US marketplaces: Google BigQuery as the core warehouse, dashboards in Looker and Power BI. Full product cost structure (procurement, marketplace fees, logistics), profitability at SKU level with the cost drivers identified, plan-versus-actual, month-over-month and cumulative metrics, and analysis of shipment routes and allocation to the nearest fulfillment or pickup points to reduce delivery cost and lead time.

B2B SaaS · subscriptions to business clientsFor a B2B SaaS company selling subscriptions to business clients: a Power BI model combining billing, subscriptions, CRM, customer activity and marketing data; a DAX metric layer for MRR, ARPU, churn, retention, LTV, plan-versus-actual and cohorts, with month-over-month, year-over-year and rolling-period logic; marketing, sales and employee KPI reporting on top; cleanup of duplicates, inconsistencies and incorrect master data; executive dashboards delivered to management.

Global distributor · powersports and recreational vehicles · 1,000+ employeesFor a global, multi-brand distributor of powersports and recreational vehicles with over 1,000 employees: ERP, CRM, service and accounting data integrated into a ClickHouse warehouse with dbt transformations and Airflow pipelines, reported in Apache Superset and Power BI. Revenue, expenses, gross margin and cost structure by product, region and business unit; plan-versus-actual by department, region and branch; stock on hand, in-transit inventory and warehouse balances across multiple warehouses and countries; bonus eligibility by plan fulfillment; warranty, service-request and repair analytics, with metrics kept consistent across business units.

Clients are not named here. US client reference available on request.

How an engagement runs

  1. I

    Discovery call

    Which decisions the reporting has to support, and which systems hold the data.

  2. II

    Data audit

    Sources, quality, gaps, duplicate metrics and the reports that disagree with each other.

  3. III

    Model

    Warehouse, transformations and metric definitions agreed with the business, in writing.

  4. IV

    Reporting

    Dashboards for owners, department heads and teams, on the model, not beside it.

  5. V

    Handover

    Documentation, definitions, refresh and monitoring in place. Your team runs it.

Start with the numbers you don’t trust.

Reply with the two numbers that never match and which systems they come from. That is enough for a first call.

Email dauren.m@lief.devEmail Dauren

WhatsApp · Telegram