Milka Santana

Abundance isn't automatic. Someone has to build it for everyone.

BuilderNonprofit founderFormer Wall StreetNew Yorker

I took a nonprofit from paper and spreadsheets to systems I designed and built myself, for finance, families, records and inventory. Each one gives leadership a live view of what's working. I bring the discipline of Wall Street, the speed of a founder, and the proof of having done it.

Milka smiling behind a row of AI robot carsBuilder
Milka at the Banco BHD Mujeres que Cambian el Mundo awardsAward winner
Milka speaking on a panel at World Woman at UNGASpeaker
Milka with the Perfecto Labs teamLeader
As seen in Forbes DominicanaEl CaribeDiario LibreListín DiarioRevista MercadoTelemicro Spoke atWorld Woman @ UNGA
How I work

Build it. See it. Lead it.

Using AI is easy. Running an organization on it means building the tools, measuring what they change, and bringing people and funders along.

Build

I design and ship the systems myself with AI, so ideas become working tools in weeks instead of quarters.

5 systems in daily use: finance, parent portal, parent messaging, inventory, student records

See

Every system is built for visibility. Leadership and the board see tuition, attendance, grades and spending in real time, not at the end of the quarter.

Quarterly board financial reviews · income statement built into the billing app

Lead

I raise the money, build the partnerships, and take the stage. Banks, tech companies and donors back the work because they can see it.

[$ total] raised · 25 staff · NVIDIA, Banco BHD, MathWorks, Altice

The transformation

Before and after AI

Drag the slider to compare how each part of the organization ran before, and how it runs now.

Screenshot comingSpreadsheets and paperTuition tracked by hand across files. Totals rebuilt every month.
Screenshot comingPerfecto FinanceOne app for charges, payments and expenses, with an income statement the board reads live.

Result: [add: hours saved per month, or how fast you close the books]

Results

From 7 students on a floor to 300

Students enrolled

Full-time school and after-school programs combined. Hover or tap a point.

See the numbers
PeriodStudents
25staff I hired and lead
55%of robotics and programming students are girls
31%MAP growth in 8th-grade math
7NVIDIA Jetson AI computers donated
[$ total]raised from banks, tech companies and donors
The journey

A New Yorker, trained around the world.

Every road leads back to New York.

The journeyStay until the end.

Every road leads back to New York. Watch where this one lands.

Stay until the end…
Selected work

The systems I built

Each one solves a real operating problem. Student names and family data are kept out of every screenshot.

Finance & operationsIn use

Perfecto Finance

Problem
Tuition, payments and expenses for about 200 students needed one reliable record, and the board needed real financial statements.
Built
A billing and accounting app with charges, payments, receipts, filters and an income statement. Future charges drop automatically when a student leaves.
Tools
ClaudeNext.jsReactSupabase
Result
[add result]
Finance & operationsIn progress

Perfecto Inventory

Problem
A technology school runs on equipment and supplies, and all of it needs to be tracked in one place.
Built
An app to manage all of the school's inventory. [add key features]
Tools
ClaudeSupabase
Result
[add once it's in use]
Parents & familiesIn use

Perfecto Student Portal

Problem
Parents had no single place to see grades, behavior and school life.
Built
A parent portal with grades, a values-and-habits points system, photo galleries, birthday notices and bulk WhatsApp messaging.
Tools
ClaudeNext.jsSupabaseWhatsApp
Result
[add: e.g. % of families logging in]
Parents & familiesIn progress

AI parent messaging

Problem
Every family needs a personal update on grades and payments, and 200 messages by hand doesn't scale.
Built
A pipeline that turns student records into a personalized WhatsApp message per family, tested side by side on Claude, GPT and Gemini.
Tools
ClaudeGPTGeminin8nElevenLabs
Lesson
My first agent setup ran into WhatsApp ban risk and rate limits, so this version is built around a model comparison and a more stable sending path.
Students & researchStudent-led

Cacao AI Lab

Problem
Monte Plata is one of the country's biggest organic cacao regions, and farmers judge pod and seed quality by eye.
Built
Students trained a computer vision model that classifies cacao pod health, and are building seed sorting and fermentation monitoring.
Tools
NVIDIA JetsonRoboflowRaspberry PiArduino
Result
Pod classifier complete. NVIDIA donated seven Jetson computers to the lab.
Growth & new venturesOct 2026

AI-built fundraising event

Problem
A Santo Domingo fundraiser for a second campus needed a full set of materials, with no design team.
Built
The presentation, posters, QR cards, invitations and social banners, all made with AI. The look became the official branding.
Tools
Claude[add tools]
Result
[add: guests, organizations, pledges]
Growth & new venturesPrototype

LatamU

Problem
Strong students in Latin American high schools are hard for U.S. and European universities to find.
Built
An AI-first recruitment tool that matches schools with universities, starting with Santo Domingo and Medellín.
Tools
ClaudeSupabaseSQL
Status
MVP scoped and dataset of 11 schools built.
On stage

Speaker, panelist, award winner

I spend a lot of my time in front of funders, partners and the press. Tap any photo to enlarge it.

Point of view

What I believe about AI

Opinions earned by building, not just by reading the news.

If you can't show what it changed, it didn't work.

Every tool I build has a number attached: time saved, money tracked, families reached. That's how I'd judge any AI investment.

Kids should build with AI, not just chat with it.

My students trained a model to sort cacao for local farmers. That's AI literacy: understanding the tool well enough to make it useful.

Privacy comes before features.

When you build for families with the least margin for error, their data is the first thing you protect, not the last.

Small organizations can transform faster than big ones.

With the right tools, a 25-person team can run with the visibility of a bank. The gap is leadership, not budget.

Career

From Wall Street to a classroom floor

2020 – nowPerfecto LabsFounder & CEO · Yamasá, DR
2021 – 2022Tubular LabsFP&A Manager · Silicon Valley startup
2016 – 2018J.P. MorganFinance Senior Associate, Consumer Banking · New York
2012 – 2016Johns Hopkins–Nanjing Center · Xi'an Jiaotong UniversityMaster's in Economics, U.S.–China International Relations, in Mandarin
2008 – 2012Goldman SachsFinance Associate, Planning & Analysis · London, then Hong Kong
2006 – 2008J.P. MorganFinancial Analyst Program, Commodities · New York
2006Baruch College, CUNYFinance and Mathematics
Let's talk

Let's build what's next.

milkasantanagalvez@gmail.com LinkedIn WhatsApp +1 917-864-3924