Notes on product building, AI, and the decisions behind the work. Written for general readers and technical leaders alike.
I used Kaggle Titanic to rebuild my machine learning fundamentals from the ground up: build a clean baseline, keep the experiment structure versioned, compare simple models, and create a submission path I can improve over time.
A small monthly subscription was enough to make me rethink the architecture of UrSalon Preview. The result was a Vercel-native stack that consolidated hosting, auth, database, storage, workflows, and billing into one operating model.
I passed the DP-700 Microsoft Fabric Data Engineer Associate exam after months of using Fabric in real revenue recovery work, then reinforcing that experience with Microsoft Learn, FabCon, YouTube, and hundreds of practice questions.
Two weeks, one complete rebuild, and a fintech automation platform that now runs on a single billing surface. Here is what drove that decision and what it actually took to ship.
The most interesting thing about agentic AI is not that it can talk. It is that it can act inside a controlled boundary, with local execution, deny-by-default policies, and human approval where it matters.
A Microsoft Fabric pipeline started as a familiar star schema, became a snowflake as cross-system relationships surfaced, then flattened back into a star for analytics performance and usability.
I had a free weekend, opened EspressoLytics on my phone, and within ten minutes had a list of things that bugged me. Here is what I fixed.
A simple request at home turned into a 48-hour iOS build and a clear founder lesson: start with one real workflow, ship fast, and use data to make decisions easier.
Over one focused week, I compared Unreal, Unity, and GameMaker, explored publishing paths, and committed to Unity after building momentum through the Essentials pathway.
AI feels trustworthy when it reduces user effort and protects identity. This is how UrSalon Preview became promptless, constrained, and transparent by design.