01.1
Analytics engineering
Models that reflect your business.
I turn operational data into tested models, with business definitions and calculations documented so you can trace each metric back to its source.
Discuss this serviceANALYTICS ENGINEERING. AI-READY DATA.
Good decisions
start with
dependable data.
I build data pipelines, analytics models, MCP integrations, and AI agents, with access controls suited to the data and the task.
ERWork directly with Etiosa RichmoreAnalytics Engineer · Malmö, SwedenREFERENCE ARCHITECTURE01-04
From source data to analytics and AI.+
From your systems
to analytics and AI.↓
to analytics and AI.↓
Data work you can check.
Traceable logicTested modelsDefined access
01 / EXPERTISE
Data engineering
and AI services.
I can help with a specific pipeline or model, or plan how your data systems fit together.
01.2
Data pipelines
Connect and maintain your data flows.
I connect APIs, databases, and business systems, and build ingestion and transformation workflows with monitoring and ways to recover from failures.
Discuss this service01.3
AI data readiness
Set the rules for AI data access.
I classify sensitive fields and define masking and access rules based on what each AI tool needs for its task.
Discuss this service01.4
MCP development
Connect AI tools to business data.
I build Model Context Protocol (MCP) servers that connect AI tools to your business data. That includes preparing datasets, defining the exposed tools and resources, and adding authentication and row-level security where needed.
Discuss this service01.5
AI agents
Automate a defined task.
I build agents for specific business tasks and define which data and tools they can use. I deploy them in an isolated environment, such as a dedicated VPS, with scoped credentials and approval steps for sensitive actions.
Discuss this service02 / SELECTED WORK
Notes from
a data project.
This example shows how a name can remain exposed after masking. I explain the work in more detail at Behind the Query.
DATA GOVERNANCE × AIWhen masking
Read the engineering notes
When masking
misses a name.
Masking the customer name field is not enough if a deal label still contains the name. The derived field needs to follow the same masking rules.
What I worked on
My work covered customer classification across historical records, warehouse masking, semantic-layer controls, and checks for names exposed in derived fields.
ILLUSTRATIVE EXAMPLESQL / LINEAGE
Masking the customer field alone
- customer_name
- Customer 1003
- deal_label
- Daniel Reyes / New deal
The derived label still contains a personal name.
Consistent protected labels
- customer_name
- Customer 1003
- deal_label
- Customer 1003 / New deal
The deal label now uses the masked customer name.
Synthetic records, reconstructed from my published explanation. These are not client data or production metrics.
03 / APPROACH
How we'll
work together.
We start with what you need the data to do and what isn't working today.
- 01
Review your setup
We review your systems, business definitions, and constraints, then identify what needs fixing.
- 02
Define the work
We agree on what I'll build, who is responsible for each part, and how we'll check the result.
- 03
Build for the handover
I test the work and document the models, assumptions, and decisions so your team can maintain it after handover.
04 / THE PERSON BEHIND R DATABLOCKS
Work directly
with me.
Etiosa Richmore Analytics Engineer
I'm an analytics engineer with a background in finance and on-chain analytics. I build data models and control how analytics and AI tools access the data.
Through R DataBlocks, you work directly with me. I write about data modeling, access controls, and the decisions behind the work at Behind the Query.
Explore my technical writing05 / START A CONVERSATION
What needs
to work better?
A pipeline to improve, data to connect through MCP, or a task to automate with an AI agent. Tell me what you have in mind.
etiosa@rdatablocks.comYour enquiry comes directly to me.