Things I've built.
Outcomes that shipped.

The Problem

Market data was everywhere. Actionable signals were nowhere. A retail intelligence company needed a way to monitor competitor pricing, inventory shifts, and market signals automatically — without drowning analysts in noise or building a 10-person data team.

Approach

Built an AI-powered monitoring system that ingests structured and unstructured market data, applies ML-based signal detection, and surfaces alerts with confidence scoring and source attribution. Designed the product from first principles around what decision-makers actually needed to act on, not just what was technically possible to surface.

Outcome

Reduced analyst time on market monitoring by ~60%. Increased signal confidence with a built-in explainability layer that let users understand why each alert fired. Shipped to first enterprise customers within 14 weeks.

The Problem

Small businesses have real customer communication needs but rarely have the budget, time, or technical staff to implement AI-powered solutions. Most tools require onboarding that takes weeks and dedicated staff to manage.

Approach

Built an LLM-agnostic SMS assistant with a dead-simple setup flow. Business owners configure it in under 30 minutes. Handles inbound customer messages, routes inquiries, answers common questions, and escalates to humans when needed. No app required for customers — just a phone number.

Outcome

In active development. Designed for zero-tech-staff deployment. Early access available for interested businesses.