Led the architecture of the Workiz Assist platform, combining Python 3.11 services, OpenAI and Azure OpenAI APIs, Snowflake, and Google Cloud Platform (GCP) to deliver production AI capabilities for service-business operations. Designed retrieval-augmented generation (RAG) pipelines that ingested knowledge articles, job records, and operational documents from Cloud Storage into BigQuery and Snowflake; generated embeddings, maintained vector indexes, and applied metadata-driven filtering for tenant, source, and document ownership. Built tool-calling AI agents for scheduling, customer communication, and field-service workflows, implementing prompt versioning, structured outputs, PII redaction, authorization guardrails, offline evaluation datasets, and human-escalation paths before releasing features to customers. Owned the migration of operational reporting datasets into a governed Snowflake data model on GCP, defining dimensional schemas, incremental ingestion patterns with Dataflow and Cloud Composer, YAML-based source configurations, and Terraform 1.6+ deployment modules; reduced duplicated pipeline logic across product teams. Established metadata and data-product conventions for the Service Intelligence data products, documenting business definitions, lineage, freshness expectations, ownership, data contracts, and operational runbooks so analysts and downstream engineering teams could discover and safely reuse curated assets. Directed technical design reviews and mentored engineers across application, data, and AI workstreams; partnered with product managers, architects, and support operations to translate requirements into scalable pipelines, investigated production data-quality incidents, and strengthened CI/CD checks for SQL, Python, YAML, and infrastructure changes.

Designed multi-tenant SaaS data services with Java 11, Spring Boot 2.7, Python 3.10, PostgreSQL, and GCP, separating tenant configuration, operational events, and analytics models to support secure product integrations and reporting. Built ingestion and transformation workflows that loaded partner and customer data from Cloud Storage into BigQuery, using declarative YAML mappings, idempotent processing, validation rules, and replayable jobs to make integrations easier to operate and migrate. Delivered the FinOps Insights platform, modeling subscription, usage, account, and transaction data into dimensional structures for finance and customer-success reporting; documented source-to-target mappings, metric definitions, ownership, and recovery procedures for each data product. Integrated Stripe and Dwolla payment workflows and Plaid account connectivity into the broader multi-tenant SaaS platform, normalizing external events into governed transaction models while handling webhook retries, reconciliation exceptions, tenant isolation, and audit requirements. Supported a modernization effort from application-owned reporting tables to BigQuery-backed analytical datasets, coordinating schema changes with analysts and platform engineers and adding CI/CD validation for SQL, YAML configuration, and deployment artifacts. Reviewed Java, Python, and React 17 changes, mentored engineers on API and data-contract design, and diagnosed production issues involving late-arriving events, duplicate integration records, and inconsistent tenant metadata.

Developed Java 8 and Spring Boot 2.x services, React 16 interfaces, and Python 3.7 utilities for order management, inventory, billing, and operational reporting within a cloud-hosted business platform. Created reusable ingestion components for CSV, REST, and relational sources, applying SQL validation, normalization, incremental loads, and audit fields before publishing curated datasets for reporting and downstream product workflows. Implemented the Operations Analytics workspace with dimensional models for customers, products, orders, locations, and transactions, documenting transformation rules and report-level definitions so business users could consistently interpret metrics. Contributed to a HIPAA-aware patient-services integration that exchanged eligibility and claims-related data with external systems, enforcing least-privilege access, encryption, audit logging, and controlled handling of patient identifiers without exposing clinical data in operational logs. Maintained deployment configurations and environment-specific YAML artifacts for application and data services, adding automated test and release checks that reduced configuration drift across development, staging, and production environments. Investigated production defects across Spring Boot APIs, React workflows, and SQL workloads, partnered with support and business teams to isolate data discrepancies, and documented operational runbooks for recurring reconciliation and pipeline failures.