Building things that solve real problems — from AI phone agents and financial dashboards to zero-UI automation tools. Over a decade of shipping software across enterprise platforms, data systems, and side projects I couldn't stop thinking about.
Originally EchoBankAgent — a unified UI platform for managing bank accounts and cards across multiple financial institutions, with voice access through Amazon's Echo platform. The entire application was hand-coded from scratch over a full year of intensive development, pre-dating the LLM era. Every single UI component, backend service, and third-party integration was painstakingly built by hand.
A data-driven analytics platform for discovering trends in home sale and rental prices across North Carolina's Triangle area. Under the hood, serious data scraping and analytics scan the price action of thousands of local properties to surface what an average home is likely to sell or rent for — accounting for highly localized variables like neighborhood, home type, home size, and moments in time.
An AI-powered team generator built for a local soccer group to solve the recurring headache of manually sorting 30+ players into balanced teams. Instead of someone tediously reviewing each player, a user simply uploads a screenshot of the player roster and the app uses AI to automatically form 2 or 4 balanced teams.
A true RAG (Retrieval-Augmented Generation) chat tool that makes HOA governance documents actually accessible. Instead of sifting through dense Covenants and Bylaws to figure out what's allowed, residents can simply ask a question and get an accurate, citation-backed answer. The model has access to all HOA documents alongside NC state HOA Bylaws for comprehensive answers.
A zero-UI calendar assistant that turns plain text messages into Google Calendar events instantly. No app to open, no forms to fill out — just text something like "Remind me to call Allen at 2:30 PM tomorrow" and the event is created on the spot. Saved as a contact called "Calendar Bot," it chains seamlessly with Siri or Google Assistant for completely hands-free event creation while driving or running.
An AI-powered phone agent platform that lets local businesses set up intelligent call handling on their existing phone numbers. The AI receives customer calls and handles call routing, appointment scheduling, FAQs, spam blocking, and phone-based form completion in real time. The key differentiator: unlike most competitors that require a new number, this works with the business's own native phone number via VoIP migration.
Modernizing legacy platforms across the organization. Building microservices for business-critical applications from the ground up, deploying cloud infrastructure, and collaborating with data architects and product owners on compliance, technical, and product delivery.
Led Wells Fargo's patch-automation framework across 20,000+ servers after the prior team's sudden departure. Separately led a zero-downtime TIBCO-to-MuleSoft migration, completing it in 9 months and later rebuilding the stack with Spring Framework.
Worked on Fusion — the search and recommendation engine behind Best Buy, Bank of America, and Reddit. Integrated spark-nlp into the core platform, built search pipeline stages, and productionized a Question Answering feature from research code.
Built end-to-end AEM web applications for enterprise clients like Ingersoll Rand, Spectrum, and K12. Led SOLR search architecture, integrated third-party services, and served as acting tech lead managing client delivery within an agile workflow.
Studied software engineering spanning requirements analysis, UML design, and implementation in Java/J2EE, PHP, and Python. Built a bioinformatics web interface on J2EE and pursued thesis research on PHP static type inference to help developers write less error-prone code.
Currently a Principal Software Engineer at Fidelity Investments. Always open to interesting conversations and collaborations.
Connect on LinkedIn ↗