Skip to content

Luca Data Solutions

AI and data engineering for HubSpot SMEs.

I build HubSpot systems and the data infrastructure around them. Integrations, automation, AI workflows, and reporting designed for growing companies.

See systemsBook a call (opens in a new tab)

What I build.

My projects usually span multiple areas. CRM architecture, integrations, migrations, AI automation, and reporting.

  • 01

    HubSpot CRM optimisation

    Cleanup, data models, custom properties, workflows, reporting layers, and pipeline architecture.

    For companies whose HubSpot instance has grown messy over time.

  • 02

    AI workflow automation

    LLM enrichment, lead scoring, classification, and operational workflows built with OpenAI, Anthropic, or HubSpot native AI tools.

    For teams that want AI integrated into real operational systems.

  • 03

    Data engineering

    Pipelines, ETL/ELT, warehouse setup, and integrations between HubSpot and external systems.

    For teams whose data lives across CRMs, spreadsheets, S3, Redshift, Snowflake, or internal tooling.

  • 04

    Data migration

    Migrations from Excel spreadsheets, legacy CRMs, or data warehouses. Source mapping, deduplication, validation, and cutover plans.

    For teams replatforming onto HubSpot or consolidating data into one system.

  • 05

    Data quality and property audits

    HubSpot property audits, data diagnostics, reporting infrastructure, and executive dashboards.

    Often the starting point with teams that no longer trust their data.

  • 06

    CRM infrastructure tooling

    Internal tools that make HubSpot schemas declarative and deployable. Defined in Excel or UI layers, synced through APIs, versioned, and reversible.

    For teams treating CRM architecture as infrastructure, not admin work.

Systems and tooling.

Proven systems designed for HubSpot.

HubSpot Property Audit Algorithm.

A scoring engine for identifying stale, risky, and removable HubSpot properties across mature portals.

Silent demo. See description below.

Problem

Mature HubSpot instances accumulate hundreds of properties, workflows, and custom objects. Cleanup becomes operationally risky without a defensible audit system.

Outcome

A ranked, reversible cleanup plan that allows non-technical admins to safely reduce schema complexity without breaking operational workflows.

Delivery

Delivered as a cleanup report. Output in Excel or the reporting tool of your choice.

Request an audit

SchemaOS.

Declarative infrastructure for HubSpot schemas.

Silent demo. See description below.

Problem

HubSpot schema implementation is still largely manual. Large builds require thousands of repetitive UI operations across properties, objects, associations, and pipelines. A 300-property implementation can mean 5 to 15 hours of manual configuration work alone.

Outcome

HubSpot schemas defined as structured configuration and deployed through a repeatable, versioned workflow. Once approved, large-scale builds deploy in minutes rather than days.

Discuss a build

Background.

Quantitative economics by training. HubSpot systems, AI and data engineering by practice.

Work that sits between CRM architecture, data infrastructure, and AI engineering, applying quantitative methods to operational systems for growing companies.

Education

  • 2020–2023

    BSc Economics

    Lancaster University, Department of Economics

  • 2023–2024

    MSc Economics, Computational

    Lancaster University, Department of Economics

    Distinction. Awarded the Prize for Overall Academic Excellence.

    Dissertation on computational macroeconomics: a non-linear real business cycle model with stochastic-volatility uncertainty shocks, Epstein-Zin preferences, and financial frictions, solved with third-order perturbation methods in MATLAB.

  • Commencing October 2026

    MSc Statistics and AI

    Lancaster University, Department of Mathematics

    Focus areas: deep learning, Bayesian neural networks, probabilistic modelling, and statistical learning systems.

Practice

CRM systems

HubSpot as the primary implementation and infrastructure platform. Migration and integration work across Salesforce, Zoho, Workbooks, Excel, Snowflake, AWS S3, AWS Redshift, and Azure Blob.

Data engineering

Python in production. SQL and warehouse modelling. ETL and ELT pipelines, API integrations, and data validation workflows. Production experience spans high-throughput operational data at a major UK consumer business and full-time data engineering inside a HubSpot agency.

AI and automation

OpenAI and Anthropic APIs. HubSpot native AI workflows. LLM enrichment and classification systems. Multi-step agentic workflows and operational tooling. Underpinned by postgraduate work in data science, advanced econometrics, and the statistical foundations of AI model design.

Quantitative and computational

Dynare. MATLAB. Numerical modelling. Simulation systems. Statistical experimentation. Computational economics.

Method

I approach HubSpot implementations as operational systems rather than isolated CRM configuration tasks.

Before building workflows, integrations, or reporting layers, I map the data flow, the operational dependencies, the failure modes, and the human decision points.

The result is infrastructure that is structured, maintainable, and understandable by the next person who inherits it.

Get in touch.

A 30-minute call is the fastest way to know if there is a fit. Email also works.

Book a call (opens in a new tab)Email instead

luca@lucadatasolutions.com

Book a call