Engineering

I have always adopted early what the market would only use years later: distributed architecture in 2009, AI classifying conversations in 2012, Domain Driven Design in PHP in 2013. I have been a CTO my whole career, and that foundation is what makes an AI agent work in production, not just in the demo.

Technical track record

2009 · ScalabilityDistributed architecture before it was the normQueues with Gearman, caching with memcached and distributed Nginx to handle volume, when most teams still scaled by buying a bigger server.
2012 · Wemind · Founder · ExitAI classifying conversations, a decade before LLMsSocial media monitoring and customer care: queued workers collecting Twitter, Facebook, YouTube and blogs, with automatic classification of replies and customer behaviour analysis.
2013 · 123i · Founder · ExitDomain Driven Design in PHPDigital real estate marketplace built with DDD, when almost no one in Brazil applied DDD, let alone in PHP.
Shopback · CTO · ExitShopbackCashback and loyalty.
Aprova Digital · CTOAprova DigitalGovTech reaching 21 million people per year. I rebuilt the architecture with queues, caching and DDD, achieved PCI certification and introduced Shape Up, observability and a 2h incident SLA.

Architecture in production

AWS & Totvs FluigSecure and efficient AWS architecture for Totvs FluigA foundation that protects the data and keeps operations running without interruption. 99.9% uptime.
AWS & CloudflareScalable infrastructure for Black Friday BrazilAutomatic scaling, storage optimisation and real time product updates to hold the peak, with zero downtime.

Open source

Need someone who has scaled this before?

Architecture, a tech team or AI in production: the conversation starts with your operation's problem.

Work with me