Real tools built to solve real problems — each one from concept to production.
AI-powered customer persona extraction platform that turns raw customer data into strategic marketing personas.
The Problem
Building customer personas used to mean locking a cross-functional team in a conference room for a week with sticky notes and assumptions — an exercise that produced artifacts nobody trusted and everyone ignored.
The Approach
Built over two and a half months using AI-assisted development tools on a modern full-stack platform. The system ingests multi-source customer data — CRM records, survey responses, behavioral analytics — and uses Claude's LLM to generate informed, data-driven personas complete with journey maps and strategic recommendations.
The Result
A flexible system that creates personas in minutes instead of days or weeks, with results grounded in evidence rather than intuition. Includes sharing, PDF/PPTX export, and team collaboration features.
Next.jsReactTypeScriptTailwind CSSClaude APIPrismaPostgreSQLStripeNextAuth
A Python-based agent suite that automates the marketing operations work machines should handle.
The Problem
Marketing operations teams spend an unreasonable amount of their time on work that machines should handle — validating UTM parameters, pulling campaign reports, watching for anomalies that signal something's gone wrong.
The Approach
Built three purpose-driven agents: a UTM validation agent that enforces naming conventions and catches errors before they corrupt analytics, a campaign insight analyzer that uses RAG to surface actionable patterns from performance data, and an anomaly detection system that alerts Slack when metrics move outside expected ranges.
The Result
Three problems. Three agents. Zero manual QA. The suite is designed for handoff to engineering teams, with clean documentation and production-ready architecture.
PythonOpenAI APIscikit-learnRAGSlack IntegrationPyYAML
A personal productivity system that turns weekly work reflections into AI-generated career summaries.
The Problem
Most professionals know they should document what they've accomplished — for performance reviews, for career clarity, for the simple act of not forgetting what they've done. Almost nobody actually does it, because the tools don't meet them where they are.
The Approach
Built on Next.js with Notion as the database backbone, the hub accepts voice or text input for weekly reflections. Claude AI synthesizes months of entries into coherent monthly summaries and semi-annual reviews, turning scattered notes into structured career narratives.
The Result
Turns the 'I should really write that down' impulse into a system that writes itself. Structured reflection workflow with AI-powered insights, accessible from any device.
Next.jsReactTypeScriptNotion APIClaude APIWeb Speech APITanStack Query
Have a project in mind? Let's talk about what I can build for you.