Breaking the Cycle — Colleagues Are All You Need

Computer Vision & Machine Learning Researcher
My work focuses on image domain transfer, anomaly detection, and generative modeling for visual understanding.
A unified platform for managing and interacting with AI agents across machines and cloud environments. It offers a streamlined interface for messaging agents, monitoring their behavior, and securely controlling and delegating tasks across distributed systems.
A lightweight watch app for "did I do it?" tasks. It keeps a visible ON/OFF reminder on your watch face, supports multiple independent toggles, and can auto-reset daily or at fixed intervals.
An occupational therapy evidence-based practice copilot powered by MedGemma. Helps clinicians move from patient questions to usable evidence—fast, structured, and citation-backed.
Interactive therapy games designed for dichoptic vision treatment. Features engaging gameplay mechanics to help patients with amblyopia and other binocular vision disorders.
How to build a layered testing system for software built with AI agents — from regression debt and test taxonomy to evaluator economy, diff-to-risk-to-scope, and cheap trust.
Read the articleWhy future-intention failures happen, when persistent external cues help, and how a simple watch toggle can support habits, safety checks, and everyday remembering.
Read the articleHow a single Gum-powered CLI keeps AI-assisted projects organized, safe, and easy to onboard—even as scripts multiply.
Read the article