AI Product Engineer · Technical Founder
AI Product Engineer and technical founder building production AI systems end-to-end — from product architecture and full-stack applications to GPU inference infrastructure and commerce integrations.
~30K/mo
Virtual try-ons
87%
Lower inference cost
36%
Lower inference latency
20+
Production stores
Featured product
AI SaaS for fashion e-commerce
Founder & Lead Engineer — end-to-end ownership across product requirements, system architecture, frontend, backend, data modeling, AI infrastructure, Shopify integrations, deployment and production iteration.
Fashion e-commerce buyers often cannot tell how clothing will look on their body or which size to choose. That uncertainty drives returns, lowers purchase confidence and hurts conversion.
I designed and built Omafit as a multi-tenant AI SaaS platform integrating virtual try-on, deterministic sizing recommendations, context-aware product discovery, merchant analytics and commerce infrastructure.
Conceptual flow from commerce surfaces through the application and AI orchestration layer.
Commerce
Omafit Application
AI Orchestration
Shopify Admin / Storefront APIs ↔ Application ↔ AI pipeline
DECISION 01
Third-party inference was too expensive and slow at scale.
I evaluated third-party API inference at approximately $0.075 per generation and 25 seconds. I designed a self-hosted asynchronous inference pipeline on AWS EC2 GPU infrastructure, bringing inference to approximately $0.0095 per generation and 16 seconds.
DECISION 02
Not every problem needed an LLM.
I built the sizing system as a deterministic recommendation engine combining body measurements, BMI adjustments, fit preferences, garment-specific weighting, fabric elasticity, asymmetric penalties and confidence scoring.
DECISION 03
Keeping AI recommendations grounded in real inventory.
Product data is retrieved and filtered before being injected into the conversational AI context, constraining recommendations to products available in the merchant's catalog and reducing unnecessary context.
Participating merchants reported an average 46% reduction in returns and 39% increase in conversion after two months of using Omafit.
Application
TypeScript · React · Node.js · PostgreSQL
AI
Python · Computer Vision · LLM Integrations · MediaPipe
Infrastructure
AWS EC2 · Railway · Netlify · Supabase
Commerce
Shopify Admin API · Storefront API · OAuth · Billing · Webhooks
Engineering
Self-hosted GPU inference, asynchronous processing, provider integration and cost/latency optimization.
TypeScript, React, Node.js and PostgreSQL across production SaaS applications.
Computer vision, deterministic recommendation systems, LLM workflows, structured retrieval and agents.
Shopify OAuth, Admin/Storefront APIs, billing, webhooks and storefront integrations.
Experience
Founder & AI Product Engineer
Built and operate an AI-native fashion commerce platform across full-stack engineering, AI infrastructure and commerce integrations.
2024 — Present
Founder & E-commerce Manager
Operated a fashion e-commerce business, developing the domain knowledge around conversion, sizing and returns that later led to Omafit.
2022 — 2025
About
I'm a Brazilian engineer and technical founder who likes turning ambiguous product problems into working systems.
My work sits between product engineering and applied AI: deciding what should be deterministic, what should use AI, how the system should be deployed, and how to make it economically viable in production.
I built Omafit after operating a fashion e-commerce business myself, which gave me direct exposure to the sizing, conversion and returns problems I later approached as an engineer.
I'm currently pursuing a B.S. in Computer Science and am open to remote international Founding Engineer, AI Product Engineer, Product Engineer and Full-Stack Engineer opportunities.
Contact
I'm open to remote international opportunities with product-focused teams building AI-native software.
Email me