Freelance BI Architect - Build an AI-First BI Platform from Scratch

Help us build the foundation for our next generation of Business Intelligence.

We are looking for an experienced, hands-on BI Architect to design and build a new AI-first BI platform from the ground up.

You will work closely with our junior BI Specialist and business stakeholders to define the architecture, establish trusted business models and build the first working version of the platform. This is not just an advisory role: we are looking for someone who can make sound architectural decisions and implement them.

The goal? A scalable and maintainable BI foundation that our internal team can confidently continue developing after your engagement.

The project

We already have an existing BI environment with Tableau dashboards covering areas such as operations, churn and revenue. These reports contain valuable business knowledge, but our underlying operational platform is changing.

Rather than simply reconnecting existing dashboards to new source structures, we want to use this opportunity to rebuild our BI foundation properly.

Our intended core stack is:

• Snowflake – central data warehouse

• Fivetran – managed data ingestion where appropriate

• dbt – transformation, testing, documentation and business modelling

• Tableau – existing reporting environment and functional reference

Our data comes from sources including MariaDB, MongoDB, event-sourced applications, APIs, Google Analytics and Zendesk.

Historical consolidation between our old and new operational platforms is not part of this project.

Why AI-first?

We don't want to build a traditional BI platform and add AI afterwards.

From the start, our architecture and data models should make reliable AI-assisted analytics possible. Think of:

• asking business questions in natural language;

• generating or suggesting dashboards;

• explaining metrics and their definitions;

• proactively identifying anomalies;

• tracing AI-generated answers back to trusted data and calculations.

AI-first doesn't mean putting a chatbot on top of unstructured data. The foundation needs to consist of clearly defined, governed business models that both people and AI can understand.

We also expect you to challenge us: where does AI genuinely add value, and where is conventional BI the better solution?

The challenge

Our source systems contain plenty of data, but technical data doesn't automatically represent business reality.

A customer record, for example, isn't necessarily an active customer. That might mean someone with at least one active product at a specific point in time – which in turn requires clear definitions for activation, cancellation, product status and historical changes.

Together with stakeholders, you'll turn concepts like these into clear, documented and testable business definitions.

These governed models will become the trusted foundation for dashboards, analysis and AI-generated answers.

What you'll do

You'll take the lead in designing and building the new BI foundation. This includes:

• assessing our source systems, existing Tableau environment and reporting needs;

• designing the target BI and data warehouse architecture;

• setting up Snowflake, Fivetran and a maintainable dbt structure;

• designing ingestion patterns for databases, events, APIs and SaaS platforms;

• translating raw data into reusable business entities, dimensions and metrics;

• establishing clear definitions for KPIs such as active customers, products, churn and recurring revenue;

• rebuilding the agreed operational, churn and revenue reporting on top of the new governed models;

• designing an AI-ready semantic layer;

• delivering a practical AI analytics and/or anomaly detection proof of concept;

• introducing standards for testing, documentation, data quality, lineage, security and monitoring;

• working side by side with our junior BI Specialist and actively transferring your knowledge.

At the end of the project, we want our internal team to understand not just what has been built, but why it was built that way.

What we expect to deliver together

By the end of the engagement, we aim to have:

1. a documented target architecture and assessment of the current environment;

2. a working Snowflake, Fivetran and dbt foundation;

3. an initial catalogue of governed business definitions and KPIs;

4. working end-to-end data pipelines and business models;

5. an initial set of operational, churn and revenue dashboards;

6. at least one practical anomaly detection or proactive alerting use case;

7. an AI-enabled proof of concept using governed business data;

8. documentation, knowledge transfer and a prioritised roadmap for further development.

What you bring

We're looking for someone who combines architecture with hands-on implementation.

You have:

• experience designing and building a BI, analytics or data platform from scratch;

• strong data architecture, modelling and SQL skills;

• hands-on experience with Snowflake and dbt;

• experience building and operating modern ELT/data pipelines;

• experience creating governed semantic or business data models;

• practical experience applying AI or natural-language analytics to business data;

• the ability to translate ambiguous business concepts into clear and testable definitions;

• strong communication skills and enjoy sharing knowledge with less experienced colleagues.

Experience with Fivetran and working with existing or legacy BI environments such as Tableau is highly relevant to this project.

Our wider environment includes MariaDB, MongoDB, APIs, event-sourced data, Google Analytics and Zendesk. Experience with some of these – or comparable technologies – is a plus.

Python, orchestration tools, data observability and BI-as-code are also welcome additions.

You don't need to tick every box. If you've built comparable platforms using slightly different technologies or approaches, we'd still like to hear from you. We're interested in the decisions you've made, the problems you've solved and how you approach a challenge like ours.

When is the project successful?

Success means we've built more than a collection of dashboards.

We have a scalable BI architecture, reliable data pipelines and agreed business definitions that are consistently used across reporting and AI.

Our junior BI Specialist understands the architecture and can continue developing it, and we have a realistic roadmap for what comes next.

The engagement

• Contract: Freelance project

• Duration: To be determined

• Availability: [X hours per week]

• Start: To be determined

• Collaboration: Primarily remote, with regular working sessions with our junior BI Specialist and key stakeholders

• Budget: [Hourly rate or fixed project budget]

We prefer to start with a short discovery and architecture phase, followed by implementation in clearly defined milestones.

Interested?

We'd love to understand how you approach a project like this.

Rather than a generic application, please tell us about one relevant BI or data platform you've designed and built. We're particularly interested in:

• your role in both architecture and hands-on implementation;

• the technology and modelling choices you made – and why;

• your experience with Snowflake, dbt and modern data pipelines;

• how you've approached AI within BI or analytics;

• how you would make AI-generated answers reliable when business definitions are complex;

• how you would approach the first four weeks of this project.

If relevant, we'd also love to hear about experience migrating or rebuilding legacy/Tableau reporting environments.

Please include your availability, hourly rate and preferred engagement model.

Your experience doesn't have to look exactly like our environment. If the challenge sounds familiar and you believe you can help us build it well, we'd like to hear from you.