00
Agustina De Adamo
Help early product ideas find their shape.
Frontend & AI Product Engineer · from Argentina, for the world

I build where ideas become product.

The gap between “we have an idea” and “users can actually use this.” I find the missing decisions, shape the flow, and build the interface — with the frontend depth to keep it maintainable.

Let's talk
01 — Who

Not a designer who codes. Not a developer who avoids product.

Working remotely Agustina De Adamo working on her laptop

I'm a frontend engineer who came up through graphic design, so I notice the things that make a product feel clear and trustworthy — hierarchy, rhythm, the empty state nobody designed, the error message that reads like an apology.

But I've also spent six years in production: complex state, APIs, legacy code, refactors, testing, conversion, frontend architecture. I'm practical. I don't ship beautiful theory that nobody can maintain.

I'm useful when things are still undefined — early products, AI tools, SaaS, dashboards, internal tools, e-commerce flows — and someone needs to bring order without slowing everyone down.

02 — What I actually do

A rough idea carries a pile of unanswered questions. I answer them, then build.

we have an idea…
what's the real flow?
empty & error states
edge cases
what's the MVP?
where does it confuse?
decide before we code
…users can use it.
03 — How I think

Questions I ask before I write code.

This is the part a junior — or an AI-generated solution — skips. It's also the part that decides whether the product is any good.

01What is the real user problem — not the feature request?
02What breaks at 10× the users or the data?
03What should be documented so the team doesn't drift?
04What would a senior engineer catch that nobody else would?
0+
Years building
production software
0
US startups
& agencies
0%
Conversion lift
(checkout refactor)
0
Years graphic
design (UBA)
04 — What I'm building

Whetstone — my own product.

An AI-graded interview trainer I'm building solo, to senior standards. Two pieces: the app, in active development, and the live waitlist that validates it. Both open on GitHub — click through.

In active development

Whetstone — the app

You write your answer to an interview question from memory, an AI grades it against a reference, and you see what you covered, missed, and should fix — progress mapped across topics as you practice.

Next.jsTypeScriptPostgresSupabaseDrizzleClaude APIZod

Built to senior standards

  • Multi-tenant security at the DB layer. Postgres Row-Level Security — 28 policies across 7 tables — not app-level filtering. Verified adversarially.
  • Decisions documented. 12 Architecture Decision Records explain the why (Drizzle over Prisma, RLS over app filtering, no RAG in the MVP…).
  • AI grading layer (in progress). Server-side Claude API with structured outputs, output repair, and an eval set to catch regressions.
Live, deployed

Whetstone — the waitlist

A conversion-focused landing + waitlist, built to validate demand before building more — and to show product thinking end to end, from hypothesis to measurement.

  • Product thinking, not just execution. A stated hypothesis, one narrowed audience, one conversion goal, a measurement plan — the qualifying question doubles as the validation instrument.
  • Proportional security. Server-side validation, honeypot, in-memory rate limiting — and a documented call to skip CAPTCHA and Redis as over-engineering.
  • Server-side form handling. Airtable token stays server-side; mobile-first, accessible, with a generated OG image and a clean console.
Next.jsTypeScriptTailwindAirtableVercelGitHub Actions
Whetstone waitlist landing page — 'You know the answer. But would it actually pass the interview?'
05 — Experience

Where I've shipped — and the calls I made.

Production work for US startups and agencies. It's their software, so there's nothing to link — but here's the situation I walked into and what I drove.

Jan-WayUSA · 2024—26
Founding / Senior Frontend
SituationA 0→1 product, mostly undefined. Goals ambiguous and changing.
What I droveTurned fuzzy goals into user flows, MVP scope and API contracts. Defined a frontend architecture that could grow, then shipped end-to-end with the founder.
ResultShipped spec → production, owned not handed off.
FirstleafUSA · 2021—24
React Frontend Engineer
SituationA large wine e-commerce platform with a checkout losing people, on years of legacy code.
What I droveUntangled the checkout states and friction points, refactored legacy code, kept it accessible (WCAG).
Result+10% conversion from the checkout refactor.
Valtech2020—21
Front End Developer Ssr. · Luxury retail
SituationA luxury retailer needed a fast, polished web app and hadn't settled the stack.
What I droveHelped pick stack and architecture (React, GraphQL, Next.js SSG), built it with accessibility baked in.
ResultA performant, inclusive app on-brand.
Arcos Dorados2018—20
Front-End Developer · McDonald's SLAD
SituationInternal admin processes were manual and slow across the region.
What I droveBuilt the web app that streamlined them (React, Redux), with Jest/Enzyme testing in an Agile team.
ResultManual admin work, turned into a tool.
06 — Not sure I'm a fit?

Ask me anything. I built this to answer honestly.

Ask Agustina

grounded in real experience
Hi — I'm a small assistant Agustina built, answering from her real experience. Ask about how she works, what she's shipped, Whetstone, or whether she'd fit your team. I'll say so when I don't know.

It uses an LLM, scoped tightly to what I'd actually say. I don't trust AI blindly — so it's grounded, honest about gaps, and points you to email when it matters.