Denver, Colorado, United States
• Spearheaded the transition of a legacy desktop application's data access layer to a RESTful API with 30+ endpoints, enabling scalable data retrieval for 100+ future client applications.
• Designed a generic resource collection framework to manage limited-lifespan objects (e.g. database connections, transactions) with automatic disposal and customizable cleanup logic, now used across a 5-engineer team.
• Led the introduction of unit testing in a previously untested codebase, writing dozens of tests and increasing code coverage by 20% in the data access layer, setting a precedent for test-driven development within the team.
• Created PowerShell automation scripts for file edits that save approximately 10 minutes per task, streamlining repetitive development work and accelerating the legacy application's retrofit process across dozens of modules.
• Led a team of 4 engineers to develop IMS Tooler, a Model Context Protocol (MCP) server providing reliable database access for LLM-driven agents, ensuring traceable data retrieval.
• Developed an AI agent with Amazon Bedrock to demo IMS Tooler at Vertafore Hackathon 2025, accelerating Vertafore’s exploration of AI-assisted workflows.
• Overcame key development challenges such as LLM hallucination during tool invocation and parameter mismatches, improving reliability and tool interoperability.