Data architecture for complex systems.

Documented and designed the integration of Salesforce Revenue Cloud sales activity with HubSpot marketing data for a US nonprofit. Analyzed how accounts, contacts, opportunities, quotes, orders, and pricing-related records were created and used in Salesforce, then defined which information HubSpot required for segmentation, campaign activity, and lifecycle management. Produced source-to-target mappings, field definitions, synchronization rules, ownership decisions, and data-governance requirements to prevent conflicting updates and inconsistent customer records between the two platforms. Supported Salesforce Revenue Cloud configuration around quoting, ordering, and pricing procedures so the integration design reflected the actual sales process rather than only transferring isolated fields. The resulting design gave the implementation team a controlled approach for moving CRM data into HubSpot while preserving the purpose and quality of the underlying records.

Designed the support and case-management data model in Microsoft Dataverse and supported the migration from a self-hosted customer-support platform to Dynamics 365. Defined the Dataverse objects, relationships, required attributes, and workflow behavior needed to represent customers, contacts, cases, communications, and customer-specific support requirements. Created migration mappings from the legacy platform into the target model so historical and active support records could be moved into a consistent structure. Implemented Power Pages for end-customer access, integrated AWS Cognito for single sign-on, and used Power Automate to provision and synchronize contact records needed for authentication. Also implemented email-to-case processing and customer-specific extensions, solving the challenge of moving to a standard CRM platform without losing the workflows and access patterns required by different customers.

Designed and built a cloud data warehouse for Boekenbalie, a Rotterdam-based online books retailer. Integrated data from 20 operational and commercial systems into a single BigQuery warehouse on Google Cloud using Python ingestion and transformation workflows. Implemented a medallion architecture with bronze, silver, and gold layers to preserve source data, standardize shared entities and business rules, and publish governed datasets for consumption in Looker. The resulting platform provides a common data foundation for operational reporting, business insights, and customer analytics.

Served as strategic advisor to the CEO of a boutique consultancy, supporting business strategy, business development, performance management, and executive communication. Translated strategic priorities into defined KPIs, supporting measures, ownership, reporting expectations, and review cadences so leadership could assess execution using a consistent framework. Developed performance and operating structures that connected sales activity, client delivery, internal operations, and business-development priorities rather than treating them as separate management topics. Prepared client presentations, proposals, and executive messaging that converted complex initiatives into clear decisions, responsibilities, and next steps. This created greater consistency between the firm's strategy, the information used to manage it, and the way priorities were communicated to employees, clients, and partners.

Designed the architecture and supporting data models for an enterprise data lake integrating IBM OpenPages governance, risk, and compliance data. The source data contained dynamic object hierarchies, security-sensitive relationships, and different interpretations of shared objects across risk, control, compliance, and model-risk processes. Produced conceptual, logical, and physical data models that documented those relationships and defined a medallion architecture in which source extracts were preserved in bronze, hierarchical and relational structures were standardized in silver, and transformed reporting datasets were delivered in gold. Designed silver-layer patterns for OpenPages object hierarchies and a two-dimensional model supporting analysis by both business organization and geographic location. Also produced current-state process and data-model documentation, object definitions, schema visualizations, export specifications, reusable templates, and example query patterns, giving data engineers and business stakeholders a common blueprint for building and interpreting the lake.

Designed and implemented an automated outbound email-generation workflow for cold-lead re-engagement using n8n, OpenAI, and Microsoft Dynamics 365. The challenge was to create outreach that was scalable but still relevant to each contact, while avoiding generic messaging and making effective use of the limited information available in the CRM. Built an orchestration workflow that retrieved cold-lead records from Dynamics 365, validated the available email and lead data, and assembled contextual inputs from the last recorded conversation, lead source, and the contact's industry. Used OpenAI to generate tailored outreach that could validate continued interest, introduce new company content, or recommend a relevant resource intended to move the lead toward a discovery conversation. Implemented the process in n8n to manage CRM data retrieval, prompt assembly, content generation, and downstream email handling as a repeatable workflow. Scaled the solution across approximately 5,000 cold contacts and achieved a 5% email response rate resulting in discovery-meeting interest.

Partnered with IBM to design an AI governance quick-start offering for small businesses that needed a practical path from initial AI use-case identification to governed model deployment. The work focused on the business processes, controls, and operating outcomes required to establish an initial AI governance capability aligned with applicable US regulatory expectations at the time. Defined the onboarding journey from AI use-case intake through assessment, approval, model deployment, and ongoing governance. Designed reusable deliverables including starter control libraries, process guides, implementation templates, and day-one operating materials that customers could adapt to their own organization. Structured the framework so businesses could establish ownership, evidence requirements, review activities, and control checkpoints without first building a large governance program. Supported sales and solution-development activities both as an IBM Business Partner and alongside IBM teams, demonstrating how IBM watsonx.governance and IBM OpenPages could support the operating model and control framework. The quick-start approach continued to be offered by business partners as an accelerated entry point for customers beginning their AI governance programs.

Led a Salesforce Sales Cloud implementation that replaced legacy CRM systems, Excel-based processes, and offline sales reporting. Defined the account, contact, opportunity, and partner-related data structures needed to create a consistent view of customers, pipeline, and resale activity. Implemented automated data-quality checks and seller reminders for missing or incomplete information, addressing the challenge that reporting quality depended on regular and accurate CRM updates. Used Salesforce audit-log activity to create a gamified adoption approach that made weekly data maintenance visible and encouraged consistent seller participation. Also developed automation that matched active opportunities with relevant previous case studies and improved partner visibility into resale-related activity, making the CRM more useful to sellers rather than only a reporting obligation.

Designed and developed a custom resource-management and workforce-forecasting platform using Python and Django. Modeled employees, skills, projects, staffing assignments, planning periods, billing rates, labor costs, estimated hours, and margin targets so operational and financial planning could be managed from the same system. Built automated weekly employee and project health checks, resource-allocation tracking, staffing-plan workflows, and forward-looking capacity reporting. Combined assignment and availability data with commercial assumptions to calculate project cost forecasts, expected margins, over- and under-allocation, and future hiring needs. The platform replaced informal staffing estimates and disconnected spreadsheets with a maintainable data model and repeatable planning process for business leadership.

Designed a scalable lakehouse reporting architecture for a Fortune 500 financial services company using Databricks bronze, silver, and gold layers. Translated governance, risk, and compliance reporting requirements into logical and physical data models, with bronze preserving source fidelity, silver standardizing and relating operational data, and gold providing aggregated reporting structures. Defined row-level access-control mappings so security inherited from the source platform could be consistently applied to downstream datasets and reports. Produced architecture documentation, sample aggregations, and reporting patterns for Cognos and other consumers, reducing ambiguity for the engineering and reporting teams implementing the model.

Built a PostgreSQL-backed data model for revenue, staffing, and demand forecasting across four legal entities. Integrated CRM opportunity data with staffing plans from a custom internal platform through ETL processes, resolving differences in entity structures, planning periods, and commercial assumptions before the data reached reporting. Modeled the relationships between opportunities, expected revenue, required capacity, staffing availability, labor cost, and forecast periods so leadership could assess financial and workforce implications from the same dataset. Used Python and PostgreSQL to prepare and maintain the forecasting data, then designed a near-real-time dashboard for revenue, cost, utilization, and capacity planning. The solution replaced disconnected planning views with a consistent model for comparing expected demand against available resources.

Provided data analytics and data engineering support to Fedora Project leadership using activity data generated through the Fedora Message Bus and exposed through the DataGrepper REST API. The underlying event data represented many different types of community activity, each with its own message structure, which made it difficult to produce consistent measures of contributor participation and project volume. Designed and built a custom Python-based extraction and transformation pipeline that retrieved event data from the DataGrepper REST API and stored the raw results as Parquet files. Developed a series of Python transformers to normalize the different event structures, derive reusable activity attributes, and produce final curated Parquet datasets for analysis. Used these datasets across multiple Jupyter Notebooks to generate dashboards, repeatable queries, and exploratory insights covering packaging activity and packager volume, Fedora Account System account creation, and contributor activity across areas such as wiki edits, COPR packaging, Koji builds, and other recorded contribution types. Structured the transformation and analytical logic so activity could be grouped by contribution area and reviewed over time rather than assessed through isolated events. This gave community leadership a more consistent and reproducible view of where contribution was occurring, how participation patterns were changing, and which parts of the Fedora ecosystem warranted further analysis.

Built a PostgreSQL-based gold layer for daily employee performance-management reporting across 60 business units. The main challenge was that operational activity, HR, and payroll systems identified the same employee differently, using customer logins, local identifiers, and corporate personnel records. Designed an identity-resolution and attribution model that linked those identifiers to a governed employee record and preserved the effective organizational assignment needed for reporting. Created repeatable ETL and onboarding patterns for customer-specific source systems, allowing new operational data to be normalized into a common model without redesigning the reporting layer for each business unit. This enabled consolidated employee, team, and business-unit performance reporting from a consistent set of definitions.

Designed and documented the implementation of a regulatory change data feed for an IBM OpenPages customer. The principal challenge was translating the vendor's external data model into the customer's existing business and OpenPages object model so that incoming regulatory content could be placed correctly and assigned to the appropriate business stakeholders without relying on manual triage. Produced the technical design documentation, defined the required OpenPages objects and relationships, and created source-to-target mappings covering the vendor feed, customer business definitions, data-placement rules, and stakeholder-assignment logic. Partnered with the data-feed vendor to coordinate the feed configuration and worked with business stakeholders to validate how vendor classifications and attributes mapped to the customer's regulatory change processes. Provided the implementation blueprint used by a development team to automate ingestion, mapping, placement, and assignment within OpenPages, creating a repeatable process for routing regulatory change data into the target operating model.

Led the implementation of an enterprise master-data-management solution using Microsoft Master Data Services. Designed governed master and reference-data models for workers, login-to-employee mappings, locations, accounts, and operational classifications used across HR, operations, finance, and business intelligence. Used HRIS information as the basis for worker master data and created mapping structures that connected operational system logins to the correct employee and organizational hierarchy. Established a hub-and-spoke governance model with central standards and distributed data stewardship, allowing business teams to maintain domain data while preserving enterprise definitions and controls. Implemented version-controlled weekly releases and trained corporate and contact-center users, providing a repeatable process for reviewing, approving, and publishing master-data changes used in reporting.

Audited a five-year-old SQL Server billing implementation used to generate customer invoices. Reconstructed the underlying business and billing rules from legacy SQL because the process had limited documentation and calculation logic was difficult to trace from source activity to invoice output. Identified structural, data-lineage, and reconciliation gaps, then redesigned the query framework to calculate billable activity at minute-level granularity. Separated the calculation into independently inspectable components so volumes, rates, time rules, and adjustments could be reconciled and updated without rewriting the entire billing process. The analysis exposed material underbilling accumulated over several years; the supporting data, financial implications, and recommended remediation approach were presented to internal executives and the client's finance and operations teams.