Projects

Challenge: The Salesforce environment suffered chronic CPU timeouts and inconsistent user experience, undermining trust in the platform.

Architecture Approach: As design authority, I refactored automation patterns and introduced framework-level controls, including trigger handling, centralized error logging, and automation bypass capabilities.

Outcome: CPU failures dropped from dozens per day (often 50+) to fewer than 10 per year, and page load times improved from 6+ seconds to around 2-3 seconds.

Challenge: Commercial teams needed intelligent, configurable sales guidance directly inside Salesforce.

Architecture Approach: I designed a metadata-driven Salesforce application integrated with Watson NLP services for intent and entity analysis, enabling behavior changes without redeployment.

Outcome: Delivered AppExchange-ready packaging, reusable AI workflows, and Dreamforce demonstrations of production-capable AI-assisted selling patterns.

Challenge: Separate service desk platforms required reliable real-time synchronization across operational workflows.

Architecture Approach: I architected real-time bidirectional APIs between Salesforce and Remedy with explicit handling for identity resolution, latency, and data consistency.

Outcome: Established stable cross-system operations and a durable integration model for long-term enterprise support.

Challenge: Business and delivery stakeholders needed faster ways to understand Salesforce metadata and automation behavior.

Architecture Approach: I built and extended AI-assisted tooling, including enhancements to Salesforce MCP server capabilities, to expose flow and configuration context in a more usable format.

Outcome: Adoption increased across PM and BA stakeholders, improving system comprehension and accelerating requirement and solution discovery.