My Personal AI Lab

Welcome to my personal workspace. Below are some projects I've been working on and systems I built in AI using various tools like Claude, Gemini, ChatGPT, Higgsfield, VS Code, and many others, these are live and not mockups: a product imagery app, a job matching pipeline, a life dashboard with a second brain, an image library that scrapes the web, and social videos. They run. Most of them use Supabase, Google Cloud, various LLMs, and all sorts of goodies.

Tolerance Studio

I created this project as a tool for creating product graphics. The way it works, it uses a Python script to cut out a product image, figures out the lighting and angles, generates a background and then pastes it back together. It also has a neat text and layout editor.

The app's layout canvas: a composed KindTail ad with headline, subhead, and Shop button dragged onto the product photo, and green 'new space' zones where AI paints more scene to hit any size.
The app's tags editor: weighted Product Spec and Creative Constants tags, with Apply brand kit and Ask AI to draft tags.
The 7-point brand inspection: shape, color, materials, details, clarity, lighting, and scene, with a sample render scored 7/7 verified.
Tolerance Studio landing: 'Build a campaign from a single photo' over a row of product renders.

JobsMatch

I created this site for a friend who was searching for a job. It scrapes LinkedIn, Glassdoor, ZipRecruiter, and other sites to find listings based on their resume and experience. First it pulls out key words from experience, then uses AI to create an algorithm to find matches and ranks them against each other.

JobsMatch dashboard with four signal cards: Resume, Preferences, Learnings, and an Outlook relevance dial.
Job Board: one ranked feed across four sites with relevance scores and AI match summaries per listing.
Resume keywords panel with extracted chips that drive the per-job qualification score.
Preferences panel with job-title chips and a free-text intent field that scores a preference-fit badge on each listing.

LifeOS

LifeOS is my personal daily workhorse and I use Obsidian to ingest data about my month. What I end up with is a glorified to-do list for all aspects of my life, a QuickBooks-type budgeter, and a big-picture life coach. Also nice is my AI reads the md files when I start new projects so it has all the context it needs to see things that I cannot.

LifeOS Real Estate tab: tab bar, a Properties and assets card, and a property to-do list.
Obsidian knowledge graph: the Second Brain node map with hundreds of linked notes.

Scraper Image Library

Searching for new and updated images was so time consuming, I created an app that scrapes the image sites I need most, tags them with AI vision capabilities and Apple Vision face detection, then organizes over 1000 images so they can easily be found. Took a few days to make but has saved me weeks.

Rapp viewer: a filterable grid of Toyota gallery images with a Model / Year / Type / Lifestyle / AI Search sidebar and keyword chips with counts.
Rapp lightbox: a full-size bZ interior shot with an AI keyword panel — night, panoramic sunroof, starry sky, forest, luxury — each clickable to filter.