PeopleOS
Toward provable payroll.
I built. Directing Claude Code against written specs. Developed further by Prada Dipa.
AI Automation Commerce
Ideas become working worlds.
AI & Automation Project Manager
I turn operational problems into working systems, and direct the engineers who build them.
Ad pauses hung on one person’s memory. Now written limits check every ad.
Primarily engineered by Prada Dipa, maintained by Hady Satya.
Synthetic ads, illustrative caps
Six systems, and who built each.
Toward provable payroll.
I built. Directing Claude Code against written specs. Developed further by Prada Dipa.
One image reorder carries to every sibling page and size.
I built. Directing Claude Code.
Every AI session's progress and cost, one window.
I built. Directing AI coding agents.
Rules flag which ads to pause.
I directed. Engineering: Prada Dipa. Maintenance: Hady Satya.
Keep-or-drop call per style, from live sales.
I directed. Engineering: Kirei Kharisma Handayani.
Shoppers searching “lobster” got products with no lobster, so the shop looked broken. I fixed search so results carry the words they typed.
2 of 4 kept. 2 never mention “lobster”.
Built by me — diagnosis, implementation, release QA.
“Before” is synthetic, in the same shape as the live store’s results: there, “lobster” brought back OH CRAB! and KEY WEST. “After” runs the store’s real filter, ported.
I build pages for a live store, where a bad release costs sales rather than a quick undo.
In code: Liquid theme sections and templates.
A theme I built from scratchI moved customers’ birthday months out of free-text notes into proper fields. It previews every change before it writes one.
In code: a metafield backfill on the Admin GraphQL API, with the dry run as the default.
Run the backfillSix commits on 19 August 2026 fixed the search page, the search-as-you-type box and the suggestions.
In code: two Liquid snippets that post-filter the platform’s semantic results.
How the fix worksThe slow-seller engine gives merchandising one recommendation per style, from live sales.
Primarily engineered by Kirei Kharisma Handayani — LinkedIn profile, opens in a new tab. I defined what it had to do, reviewed the implementation, and took it through QA and rollout.
Run the engineWhen the store and the apps around it disagree, I find out why.
In practice: Admin GraphQL query cost, OAuth token lifetimes, and app figures that disagree with the store’s own data.
The roles behind itI owned releases and store checks: verify against real orders, and write down what was deliberately not changed.
Three years on one storeMy employer’s live Shopify store, not mine: what I implemented, led and directed.

Team direction
Engineers build, and so do the AI tools I direct. I set requirements, review, direct testing and rollout.
Synthetic example
Update the weekly stock sheet from the sales export.
Live
In this example, review blocks a change that would erase the sheet’s column titles.
Off the clock
Tennis, football, the gym most days, and films, watched closely. @jdenandro on Instagram — 12.1K followers, 8.3M views in the last two years. See the reels