3 spots for new projects this quarter
I'm Santiago Serrano. I listen to how your business runs, find what can be optimised and build the complete platform: data, payments, integrations and the app your team uses. And I stay with you until it's running in production.
/ I've built platforms for
/ Projects
Each one started with a conversation about how the business worked. This is what I built and what changed.
Client · MyTicket MX
/ What changed
Since September, every sale flows into its portal on its own. Loading five years of history went from 41 minutes to 6 seconds.
Client · Elements Method
/ What changed
Card and bank-transfer payments, client records and admin in one bilingual platform, live in production.
01 Origin
CRM02 Quoting engine
Multi-insurer quoting engine03 Insurers
Insurer AInsurer BInsurer C04 Result
Coverage comparisonClient · Fleet Solutions México
/ What changed
The sales team compares coverage from several insurers on a single screen, in real time, instead of quoting portal by portal.
Client · The Healing House
/ What changed
The brand sells, books appointments and publishes its blog on its own, in Spanish and English.
Client · VIP Cleaners
/ What changed
Every pair is received, charged and tracked in one system, and customers follow their order online.
01 Sources
MetaTikTokAgency file02 Download
Daily download of the last 28 days03 Structure
Common data structure (17 tables)Configuration per brand04 Results
Analysis rulesPanel per brandClient · Júpiter Studio
/ What changed
Campaign results from several brands, from Meta and TikTok, in a single data model that refreshes itself every day.
/ Try it yourself
It's the hardest problem in selling tickets online, and this is how I solved it for MyTicket. This is the real floor plan of Teatro Aldama, all 882 seats. Hold a seat, start a race between two buyers or release 40 at once: the same seat is never sold twice.
A simulation in your browser of the rule that runs on the real server: holding a seat is a single operation the database either accepts or rejects. Never both.
See how MyTicket works/ How we work
No surprises: you know what happens at each stage, what you get and when.
2–3 months
from first conversation to production in a typical project
01 / 06 — We talk
You tell me how your business runs today: what's done by hand, what gets lost, what eats your time. I listen first and propose after.
You getA clear idea of what's worth building
Before committing to the whole project, I build a version that proves the solution works with your real case. You decide with something in hand, not a slide deck.
You getA working proof and an approved plan
Together we write down what the system does, what it doesn't, and how we'll know it's done. It's the project's technical agreement.
You getRequirements document (SRS)
I work with a Kanban board you can see and a delivery calendar. Every step is tested in a real environment, not in screenshots.
You getProgress you can use and review
Domain, servers, real payments, migrated data and support on launch day. The project ends when your team is already using it.
You getYour system running with real customers
A monthly plan for improvements, servers and support. The code, accounts and access are yours from day one.
You getA system that keeps growing with you
/ Artificial intelligence
In the way I work and inside the products I deliver. Always with human review at the end.
$ claude — director
How I work
I work with ten role-based agents (architect, builder, QA, reviewer, security…) and I direct them. Every change goes through tests, review and a gate before it's merged. AI speeds things up; the decisions are mine.
In your product
Assistants connected to your system that answer on your site or on WhatsApp with real data: availability, orders, status. Like MyTicket's «Acomodador», with 12 tools on the Claude API.
↳ Illustrative example
Which campaign had the best cost per result in September?
The launch campaign on Meta: its cost per result was the lowest of the month. Here are the three sources I used. [1][2][3]
In your data
RAG systems that search your documents and databases and answer citing where each fact comes from. The rule: AI only narrates what the data has already verified.
↳ Illustrative example
/Verifiable results
automated tests passing in MyTicket
They include situations where several people try to buy the same seat at the same time.
See the casesystems already used by real teams and customers
Elements Method, The Healing House, VIP Cleaners, the Fleet Solutions platform and MyTicket's internal tool.
See projectsElements Method: 55 tables and 86 pages
Executive-coaching platform in production, with payments, content and internal tools.
Visit elementsmethod.comto load five years of MyTicket sales
45,125 orders since 2021. I redesigned the import until it stopped being a chore and became a click.
See the caseaccuracy in a computer-vision model
Recognises traffic signs; tested on more than 12,000 images it had never seen.
See Machine Learning/About me
Wherever I go, I talk to people about how they work: what they do by hand, what gets lost, what eats their time. Almost all my projects have come out of those conversations.
More than programming, I like creating solutions. What drives me is seeing something I built solve a real problem: a ticketing company selling without middlemen, a coach getting paid online, a team quoting in seconds.
Today I'm Tech Lead at Fleet Solutions México, I work in Nestlé México's digital area and I study Computer Science and Technology at Tec de Monterrey (class of 2027). Alongside that, I build platforms for companies.
I work end to end: understand the business, design the system, build it, test it and take it to production. The code and access always stay in your name.
Santiago Serrano
Software development · Tech Lead
/Contact
3 spots available this quarter
Tell me what's eating your time at work. The first conversation is free, no strings attached.
I reply personally, usually the same day.
Message me on WhatsApp