Senior leader with 13 years transforming finance through technology, including 7+ years architecting enterprise data platforms and AI-driven solutions. Recognized internationally for innovative solutions that have been adopted across multiple continents at Marriott International. Deep technical expertise spanning from low-level systems programming to cloud-scale architectures and shipped desktop software, with proven mastery of Python, SQL, TypeScript, Microsoft Fabric, and Azure infrastructure. Leading a team of 3 data professionals while driving strategic initiatives that have saved 26,000+ hours annually and scaled to serve 15,000+ users globally. Practical, hands-on approach to problem-solving with a track record of delivering production-grade solutions that scale.
Professional Experience
Key achievements and relevant contributions highlighted
Architect and own the EMEA data-analytics platform end to end: an Azure PostgreSQL warehouse (1B+ rows) with Python / FastAPI ETL pipelines feeding continent-wide analytics. I favour open-source, inspectable components, which are typically better battle-tested against vulnerabilities than closed alternatives and free of vendor lock-in.
Led the platform’s migration to Microsoft Fabric and its move away from Tableau (the US-HQ preference): chose Fabric / Power BI for lower cost, accessibility to visual developers and native Microsoft integration, and proved it with working models rather than slideware. The architecture is now reused as the global reference across EMEA, APAC and the Americas.
Designed, built and launched Kairos, the EMEA Payroll Spread Tool: a production Electron / TypeScript desktop application, live across the continent at 500+ properties, replacing the Excel / VBA payroll tool I originally launched in 2019. Its bespoke spread engine both time-phases payroll values across accounting periods and allocates them down the departmental and Chart of Accounts structure from a single input, so the calculation is defined centrally once rather than hand-rolled in hundreds of spreadsheets: that is where the accuracy gain comes from. Access delegation lets a planner distribute the work of building a plan to the colleagues who own each piece, bounded by the access model rather than by shared credentials.
Delivered Kairos end to end in three weeks as sole developer, on an offline-capable encrypted local store with a bespoke sync layer to the FastAPI back end, a hardened IPC architecture and built-in auto-updates that keep the continent on one version. I specified the architecture, data model, spread algorithm and threat model, then directed and reviewed the implementation.
Built the tool’s reporting on the same architecture as the Fabric warehouses, so measure definitions are reused rather than reimplemented and a figure inside the desktop client cannot silently disagree with the same figure in Power BI: one semantic definition served across warehouse, semantic layer, web portal and desktop client.
Engineered the security model for payroll, the most sensitive data in the estate: company SSO (Entra) brokered into short-lived signed tokens so the client never asserts its own identity, two-tier authorisation separating identity from a device-verified session, and TPM hardware binding in which the machine signs a fresh single-use nonce on every login and refresh, so a copied credential is useless off its original hardware. Rotating device-bound sessions with reuse detection and no raw secrets at rest, with Row-Level Security enforced server-side from the same access model used across the platform.
Built a self-service access-governance portal (Astro / React on FastAPI) replacing helpdesk-driven access administration: users request property, tool and data access, enrol and manage their own devices and sessions, and administrators work approval queues, device provisioning and a full audit trail. Authorisation is a first-class subsystem with an explicit rules engine and a security test suite covering scoping, device approval, TPM binding and integrity, run through CI on every change.
Designed and deployed a self-validating, hash-signed Row-Level Security and ownership-provisioning framework on top of IT’s access layer, built within Microsoft constraints (SharePoint lists and Power Platform), because the existing org-chart governance routed approvals to managers who often did not own the data. Scaled from 0 to 6,000+ RLS users (15,000 total) in two months with auditable, owner-driven, least-privilege access.
Lead and develop a multidisciplinary data team (senior manager, senior analyst, junior analyst), setting the architecture standards and delivery practices that others have adopted, and delivering hands-on training across EMEA, APAC and the Americas.
Outcome: 26,000+ hours saved annually and a 3-FTE efficiency while improving data quality and timeliness; awarded an equity grant for transformational impact.
Selected the Fabric warehouse and semantic models as the hot path for analytics and reporting over ClickHouse: although ClickHouse can be tuned faster, Fabric was fast enough, integrated far better with the existing Power BI and PostgreSQL estate, carried a Microsoft discount, and had a far lower barrier to entry. That barrier was decisive for adoption: ClickHouse would have been beyond the reach of other business disciplines and would have prevented a single, shared continental reporting source.
Brought software-engineering discipline to BI by managing Fabric / Power BI assets as code (Tabular Editor, ALM Toolkit, the Power BI project format, GitHub Actions). This made semantic models highly reusable: I redeployed a complete budget and planning reporting stack in two weeks by rotating out the sources and calculations while measures, metadata and relationships carried over unchanged, recovering substantial developer time and cutting duplication.
Built an in-house reporting portal (Astro and React / MUI on a FastAPI back end) at leadership’s request rather than relying on off-the-shelf BI, for control over embedding, report discovery and integration with the data layer. Its website experience markedly increased engagement from senior leadership (VP level and above), drawing them into the datasets; the goal was depth of executive engagement rather than raw user numbers.
Modernised legacy VBA tooling with a bespoke authentication layer, auto-updates and FastAPI integration for controlled, stable data submission.
Applied Google Material Design principles to enterprise reporting, establishing modern design standards for data visualisation across the organisation.
Manager, EMEA Financial Reporting & Analysis
Marriott InternationalJan 2021 – Apr 2023
Established the Azure PostgreSQL data-warehousing platform from scratch, scaling to 1B+ rows with multi-source ETL across continental disciplines. Chose PostgreSQL over more abstracted, managed options for cost, speed and agility: as the most widely used database it integrated cleanly with everything, and testing showed it was exceptionally fast when properly configured, letting us run mass recalculations in-database (notably foreign-currency recalculation) and pre-compute once into materialized views. Moving reporting off files and into the database brought cycles forward by around two days, enabled more tangible analytics, and future-proofed the estate.
Built a driver-based Budget Modelling Tool consolidating multiple legacy systems, generating complete zero-based budgets (5,000+ lines) in under five minutes.
Enforced structure at the atomic level by the database/warehouse (structured data) so most data lands in a silver state on ingestion; the few bronze raw sources, such as guest survey data, are canonicalised by custom in-house Python tooling that classifies sources and normalises values such as scores into measurable, cross-relatable analytics. Tuned cadence to scale-to-value: daily revenue pipelines and monthly full P&L, delivering daily revenue data where the prior file-based process managed only weekly flashes, while deliberately avoiding daily P&L where costs are not yet loaded.
Re-architected the FR&A EMEA data ecosystem (Python, VBA, KNIME, Power Query) to enable self-service analytics and significantly greater reporting depth.
Transformed manual PowerPoint executive reporting into automated Power BI dashboards; implemented continent-wide Balanced Score Cards with hotel-level drill-through and AVP / COO roll-ups across 500+ properties.
Launched an EMEA-wide Payroll Tool (Excel / VBA) adopted by 500+ properties in year one and later mandated continent-wide for budget and forecast cycles; it ran for seven years until I replaced it with Kairos, the desktop application above.
Spearheaded Power BI adoption across the continent, shifting an Excel-based reporting culture to modern, automated BI.
Modernised legacy VBA tooling across the continent, eliminating critical defects and improving performance by orders of magnitude.
Finance Analyst, EMEA Financial Reporting & Analysis (the transition point)
Marriott InternationalFeb 2018 – Jan 2019
Joined Marriott’s EMEA corporate function (initially on a temporary contract) to resolve the hardest data-merge edge cases from the Starwood acquisition, the largest merger in hotel history.
Built the EMEA PS Loader, an automated ETL tool that mapped and validated financial data from external GL systems into Marriott’s platform across Month-End, Budget and Forecast cycles, well received and copied globally; this is where my pivot from finance into data analytics began.
Architected an account-mapping framework spanning ~35 Charts of Accounts during Starwood integration; led the EMEA transition to a new CAPEX system (TREX).
Assistant Director of Finance
Baglioni Hotel LondonApr 2017 – Feb 2018
Ran full property finance (month-end close, P&L, balance-sheet reconciliation); partnered with the Italian head office to implement modern finance systems and automation; led new F&B POS, property-management and cellar-management implementations.
Finance Supervisor
The London EDITION (Marriott)Mar 2016 – Apr 2017
Supervised the finance team and built pioneering automation that became templates for other properties; ran month-end and purchasing systems; led data-driven critique meetings with the GM and department heads.
Early-Career Finance Roles
Various hotel properties, London2012 – 2016
Progressed through General Cashier, Revenue Control, Accounts Payable, Cost Control and Accounts Receivable, including an acting-supervisor role, building end-to-end hotel-finance domain depth.
Created the first hotel eDCP (electronic Daily Circulation Pack), a paperless income-audit system later recommended to other properties following an ISRA audit; built semi-automated cash-declaration and reconciliation tools (Opera vs Micros / Symphony).