Chronological updates and behind-the-scenes of current development projects.
A reflective dev-log on successfully deploying a multi-component service using modern cloud infrastructure and AI code generation, highlighting the challenges of managing LLM context efficiency and navigating the psychological shift from developer to entrepreneur when pricing a product.
The technical and strategic journey of transforming a localized, BYOK WaterlooWorks job-matching Chrome extension into a production-grade, monetized SaaS application optimized for AI agent development workflows.
Transitioning a student utility tool into a monetized product by introducing a credit system, handling sensitive data liabilities, and moving from 'vibe coding' to a structured ADR architecture for better AI agent alignment.
A reflection on building a Waterlooworks Chrome extension in two days using Claude Code, highlighting how 'harness engineering' provided full control over AI agents and drastically improved the development experience.
Introducing ww-extension, a new Chrome extension that uses AI to automate WaterlooWorks job triage, and a reflection on shutting down the unfinished Macro-dashboard project to focus on this new tool.
Reflecting on the successful deployment of the macro dashboard, the specific phase-based workflow used to guide Claude Code through the implementation, and the upcoming challenges of operations and monetization.
Starting a new project: a highly customizable finance dashboard that helps users build the habit of 'looking around' by showing asset prices and news on a single screen.
A first experience using GenAI API to extract questions from TOPIK exam PDFs. Built an extraction program with an LLM coding agent and automated the process with a bash script.
A comprehensive test of the blog's UI, showcasing various Markdown features including tables, code blocks, and typography.