Engineering profile

I manage product engineering on a loyalty platform used by 13 million people.

I work across backend, frontend and cloud, with most of my time on the backend. I rebuilt the jobs that send campaigns, evaluate audiences and run customer automations, and I handle the AWS and database work that keeps them online and affordable. On the product side I work with clients, PMs and the CEO to pin down what is actually needed, then build it end to end, as with Masivo's AI assistant and its ticketing and point-of-sale features. Before Masivo I led a backend team of twenty at Immer.

13M+people on the platform I lead technically

10xnotification capacity, 150K to 1.5M an hour

54h to 41mfor the slowest campaign workflow

3promotions at Trade EC, from associate to manager

ProductAdmin, ticketing, AI assistant

ProcessingCampaigns, audiences, journeys

DataLarge tables kept online

OperationsCost, deploys, incidents

Production results

Work I shipped and what it changed

Delivery150K to 1.5M an hour

I redesigned how campaign sends are batched, retried and rate-limited. Masivo now sends more than 3M push and email notifications a day.

Speed54 hours to 41 minutes

A campaign workflow that ran for more than two days now finishes in under an hour. Users see the same result they did before.

ResourcesAbout 80% to 4%

The same rewrite stopped workers from re-reading rows they had already handled, so they run on a small fraction of what they used to.

Audiences3M customers an hour

I rebuilt segmentation around reusable customer traits. It now gets through more than 10M records a day.

AIAssistant in production

I shipped Masivo's AI assistant. It reads each brand's live analytics and suggests campaigns or rewards, and every model call stays small and cheap.

Uptime700+ GB, no downtime

I cleaned up the largest table in the product and recovered its invalid indexes while customers kept working.

Cost37% lower AWS bill

For Metro de Quito I changed the architecture and resources and kept availability at 99.9%.

ProductQuestions before design

At Trade, requirements come from clients and from PMs, and they often arrive half defined. I send questions back until it is clear what is actually needed, and sometimes I work out product direction with the CEO. At Immer those decisions were made with the CEO.

Full stackBackend, frontend and cloud

Along with the backend and AWS, I build the Next.js interfaces: Masivo's admin, its ticketing and point-of-sale screens, all three TickPass products and the clinic workspace.

TeamsUp to 20 engineers

At Immer I led up to 20 backend developers and the frontend team on a new backend I designed as ten-plus modules.

Technology

Tools I have used in production

Frontend and product

TypeScriptJavaScriptReactNext.jsTailwind CSSReact Hook FormZodtRPCRadix UIshadcn/ui

Backend and APIs

Node.jsNestJSExpressRESTGraphQLPythonFlaskQueuesSchedulersMulti-tenant SaaS

Databases and analytics

PostgreSQLSupabaseRedisUpstashMongoDBFirebaseMySQLTinybirdClickHouseAthenaParquet

AWS and delivery

S3SQSSESWAFCloudWatchCloudTrailGlueLake FormationLambdaCloudFormationIAMSNSEC2EBSVercelGitHub ActionsCI/CDLinux

Production AI

MCPAI SDKOpenAIClaudeTool callingStructured outputsAI agentsOllamaOpenHands

Testing

PlaywrightVitestReact Testing Librarypytest

Profile sections

The rest of the engineering profile

Engineering notes

Two write-ups on Medium

AWS migration assessmentWhen the right Graviton recommendation is "not yet"I inspected a server's disk offline and found it had to be split up before moving to ARM.Agents and remote executionAutonomous coding with Ralphie, OpenHands and OllamaAn unattended coding run on Vast.ai, with repository memory and 48 external tests to check the result.

Availability

I am looking for a senior product or backend role.

I do my best work when I handle backend, frontend and cloud on the same product and help decide what gets built with clients, PMs or the CEO. I work remotely from Argentina, in English or Spanish.