JEFF KENT BUENAFLOR — METRO MANILA, PH
I am the Chief Technology Officer at Moneybees, a BSP-licensed virtual asset service provider. I spent fifteen years writing production code by hand — 1.3 million lines at Moneybees alone — and then built an AI-augmented engineering model that ships at 29 times my old team's pace, with a named human on every gate.
01 — THE NUMBERS
Every figure below is measured from the git history and pull-request records of the Moneybees engineering estate — not estimated.
02 — THE ROAD
Designed interfaces in Photoshop and shipped them in HTML, CSS and JavaScript. Learned that the pixel and the code belong to the same person.
Early AngularJS and React, back when picking a framework was a bet, not a default.
Owned the interface end to end and crossed into Node.js on the backend.
Wrote the majority of the platform: AngularJS front, Node.js and PostgreSQL back, deployed on AWS with ECS and CodePipeline.
Joined as the engineering lead of a BSP-licensed crypto exchange. Carried the platform through two generations, delivered the AFASA regulatory engineering program on deadline, stood up a three-account AWS estate under BSP-approved outsourcing, and became CTO in 2026 — the year I turned the whole delivery model AI-augmented.
03 — THE MULTIPLIER
In 2026 I designed and ran an AI-augmented delivery model for Hive 2.0, Moneybees' ground-up platform rewrite: a swarm of AI workers building in isolated worktrees, an independent AI guard reviewing every change, and a named human — me — holding merge authority alone. Proactively disclosed to the central bank.
2,709 pull requests merged in sixteen weeks — 94 percent of what the whole team merged in the previous nine years combined.
04 — THE MACHINES I LEFT RUNNING
A platform a lean team can hold has to run itself. These are systems I built for Moneybees that work while nobody is watching.
AI workers, guard reviewers and an operator loop take a ticket from validation to a reviewed, CI-green, QA-ready pull request — every artifact left behind for auditors.
runs on every ticketProves every production backup vault holds a fresh, restorable recovery point — and wakes me on the phone when one goes stale.
daily, 30h freshness gateEvery defect and pentest finding becomes a named, re-runnable guard class, so a bug fixed once stays fixed on every later change.
on every mergeCoherence probes, deploy smoke tests, gap detectors and watchdogs sweep the estate and speak up only when something is wrong.
hourly sweepsGmail, Docs and Drive automation against the raw HTTPS APIs in pure-stdlib Python — no client libraries, nothing to break at 3 AM.
on demand05 — STILL AN ENGINEER
The same git history that measures the swarm measures me. Before and beside the AI era, I remain the largest human contributor in company history.
Measured from bare clones of every Moneybees repository: non-merge commits, with lockfiles, vendored and generated bundles excluded from all line counts.