PropioHola, Rodrigo Instead of telling you how I work, I built this.

What I shipped in 60 days

Four production systems moved forward at once — institutional AI, Mexican fintech, education, and client delivery. This page is the evidence behind my answer to Propio's AI-first engineering role.

Convince me Angel
See the evidence
60days · May 21 to July 20
4production projects shipped
993commits by Angel
19.4BClaude Code token volume

“AI-first” is easy to say. The useful question is what it produces.

So instead of asking you to take the label on faith, this page shows the shipped systems, the activity behind them, and their results.

I'd rather show you the receipts.

  • 4 production projects across AI, fintech, education, and agency delivery.
  • 993 commits written by me across the same 60-day window.
  • 19.4B Claude Code tokens used as development-volume infrastructure.
  • 178.6M direct input + output tokens before cached context reads.

Four systems shipped

Each dossier starts with the business outcome and the judgment behind it. Even the videos are made with AI. I used hyperframes.

01Vera

267 commits · 60 days

AI that institutions can trust, not just admire in a demo

Vera is the AI-powered validation platform I designed and built from scratch. In this 60-day window, I kept shipping across extraction, deterministic validation, multi-tenant architecture, and the operational reality of institutions that cannot afford untrustworthy answers.

The judgment call

I did not let the model become the source of truth. AI extracts and proposes; deterministic rules verify. That boundary is what turns a fast demo into a system an institution can trust.

What reached production

  • The system cuts enrollment-document review from weeks to hours
  • AI handles extraction; deterministic checks create a trustworthy, auditable result
  • The same multi-tenant product now serves an institutional client: Colegio de Bachilleres
Open the shipped product
The resultWeeks → Hoursproduct outcome
Commits
267
Claude Code volume
5.7B
Repositories
1
02Recupera

267 commits · 60 days

One guided flow through Mexico's tax complexity

Recupera started as a side project among friends — I build it with a software engineer, an accountant, and a marketer. I own the product: in this window I shipped across the complete surface, from frontend and backend to SAT integration, documentation, scraping, and supporting tools. A simple consumer experience has to sit on top of a slow, tedious and complex SAT website, sensitive financial data, and a tax process most people would rather avoid.

The judgment call

The hard decisions were which SAT complexity the product should absorb, where SAT's slowness and failures must stay visible to the user, and which financial assumptions could never be hidden. Generating more code was the easy part.

What reached production

  • Direct SAT integration automates invoice retrieval instead of asking users for manual uploads
  • The product guides nómina employees through a declaration designed to maximize their refund
  • Eight active repositories show end-to-end ownership, not a narrow frontend contribution
Open the shipped product
The resultSAT → Refundone guided flow
Commits
267
Claude Code volume
3.7B
Repositories
8
03Prepa IN

331 commits · 60 days

A better product and a growth engine that compounds

This was the highest-velocity chapter of the window. As CTO and co-founder, I shipped across the student platform while building an autonomous content-marketing agent that researches, writes, and publishes SEO content. The job was to improve the product and build the system that helps people find it.

The judgment call

I used the agent for scale, not permissionless publishing. Research, generation, and distribution needed explicit boundaries so speed could compound without turning the brand into unreviewed model output.

What reached production

  • A rebuilt student web application and supporting API work across the GrupoLiber platform
  • An autonomous agent turns research into published growth content without a manual production line
  • The result: 21x more Google impressions and +121% organic sessions year-over-year — June 2026 was the site's best organic month on record
  • Non-brand search clicks (people who didn't already know the brand) grew 13x in a year, moving average position from 19 to 7
Open the shipped product
The result21xGoogle visibility
Commits
331
Claude Code volume
7.7B
Repositories
7
04HUSL Digital

128 commits · 60 days

AI-first delivery that holds up under client pressure

HUSL is a different kind of proof: the workflow has to survive outside my own product roadmap. Across agency and client work, I used AI-first development to compress standard builds from roughly 40 hours to 8 while still shipping against real requirements, reviews, and deadlines.

The judgment call

I standardized the repeatable parts without forcing every client into the same answer. AI accelerates the common work; the implementation still has to respect the client's system, constraints, and quality bar.

What reached production

  • Up to 80% less development time on standard builds without lowering the quality bar
  • Repeatable workflows used across client delivery, not a one-off personal experiment
  • Production work spanning multiple client codebases and two Claude Code environments
See HUSL Digital
The result40h → 8hstandard site build
Commits
128
Claude Code volume
2.3B
Repositories
12

Four parallel workstreams. One repeatable operating system.

Across different domains, the pattern stayed consistent: understand the constraint, hand the mechanical work to AI, keep ownership of the decisions, and stay accountable for what reaches users.

60days
4production projects
993commits
178.6Mdirect AI input + output
01
Vera

Weeks → Hours · product outcome

267 commits

02
Recupera

SAT → Refund · one guided flow

267 commits

03
Prepa IN

21x · Google visibility

331 commits

04
HUSL Digital

40h → 8h · standard site build

128 commits

Claude Code reported 19.4B in total workload volume, including cache reads. Direct input and output are broken out separately so the activity numbers stay honest.

The numbers show velocity. The decisions show how I work.

AI only pays off when the engineer stays responsible for context, boundaries, verification, and the production result.

  1. 01

    Set the direction

    I give Claude Code the business constraint, architecture, security boundary, and definition of done. The model gets context; I keep responsibility for the direction.

  2. 02

    Delegate the heavy lifting

    I use agents, workflows, and long-lived context to compress research, implementation, debugging, and review, then intervene where product or technical judgment changes the answer.

  3. 03

    Own what ships

    Tests, security, observability, edge cases, and user feedback close the loop. A commit is activity. A working system that changes a business metric is the outcome.

Propio

This is the execution pattern I would bring to Propio's next 60 days.

You're building an AI-first financial product with a small team and a high trust bar. Recupera proves I can turn Mexican tax complexity into a consumer product and work alongside accountants. Vera proves I can make AI useful inside an auditable institutional product.

These 60 days prove the execution loop: understand the business constraint, hand Claude Code the heavy lifting, challenge its output with my own judgment, and own what ships. That is the loop I would bring to Propio's ledger, tax-planning workflows, and the product decisions around them.

Give me Propio's hardest 60-day problem.

You've seen the outcomes, the work behind them, and the judgment I keep in the loop. Now I'd like to understand the problem that matters most on your roadmap.

Talk about the next 60 days