gxstudio.dev
Currently accepting select engineering & advisory contracts

I rescue fragile codebases and ship software that survives production.

When your Cursor, Lovable, or v0 prototype hits the scaling wall, or when you need a senior Tech Lead to architect custom software without full-time agency bloat.

Verification Latency
10m → 50s (12x faster)
Enterprise Refactor
99.9% Migration Uptime
Leadership Scope
9 Engineers / 2 Teams
AI Workflows
MCP & Claude Code
01 / The Technical Problem

Prompting gets you 80% of the way in a weekend.The remaining 20% breaks your business.

AI coding tools generate plausible code, but they don’t understand concurrency, connection limits, multi-tenant isolation, or background queues. When early users arrive, subtle architectural failures emerge:

Synchronous Execution & 504 Timeouts CONCURRENCY

AI code almost always executes long-running operations (PDF generation, webhooks, LLM calls) directly inside the HTTP request loop. When 5 concurrent users trigger it, your web workers exhaust and the app freezes.

Unbounded Queries & Missing Indexes DATABASE

Models create clean-looking ORM models without foreign key indexes or eager loading. At 50 records it is instantaneous. At 5,000 records it generates N+1 queries that lock your database.

Multi-Tenant Data Leaks SECURITY

Missing row-level security or authorization checks on endpoints. Tenants can guess incremental IDs and inspect other organizations' sensitive data.

Deployment & Operational Fragility DEVOPS

Works on localhost, crashes on cloud. Missing health check retries, unhandled environment variables, and bloated containers that cost $300/month for no reason.

02 / Case Files

Selected Systems & Architecture

CASE_01 Read the full case →

Enterprise Global Monolith Refactor & Verification Acceleration

Ruby on Rails • Java Spring Boot • Azure
The Bottleneck

Inherited a years-old legacy monolith serving global enterprise brands. Profile verification and document generation took 10 minutes per request, frequently stalling web servers and locking records.

The Architecture

Decoupled synchronous generation into an idempotent async job pipeline. Led phased 11-month transition to modern services while maintaining unbroken production availability.

Verified Impact
• 10 min → 50 sec (12x latency reduction)
• 99.9% platform uptime sustained
• Zero service interruptions
CASE_02

B2B QA Regression Management Platform — Multi-Tenant Test Engine

Rails 8.1 Solid Stack • Hotwire • Kamal VPS
The Problem

Engineering teams needed a multi-tenant test-case management system with evidence capture and bidirectional Jira sync, without the memory footprint of heavy JavaScript build pipelines or costly cloud runtimes.

The Architecture

Built on Rails 8.1 with Turbo 8 and Stimulus. Replaced external Redis with the native Solid stack (Solid Queue, Cache, Cable). Zero Node.js, zero Webpack. Multi-tenant subdomain isolation ({tenant}.domain.com) deployed via Kamal to low-cost VPS.

Verified Impact
• 0ms Node runtime footprint
• Bidirectional Jira API sync
• Ultra-lean ops on $20/mo VPS
CASE_03 Read the full case →

Clinical Healthcare & Practice Management — Split-Payment Escrow Architecture

Microservices • Rails API • PostgreSQL • Docker
The Problem

A clinical mental health platform required confidential scheduling, dynamic practitioner matching, and legally compliant financial split-payouts between clinics and therapists.

The Architecture

Architected decoupled services for Auth, Scheduling, and Financial/Billing. Built automated escrow calculations, appointment calendar locks, and profile verification pipelines.

Verified Impact
• Automated multi-party split payouts
• Sub-second schedule matching
• Strict HIPAA/LGPD compliance
CASE_04

B2B EdTech Enterprise Platform — Production LLM Tooling & TDD Overhaul

Ruby on Rails • React • HubSpot • LLM Pipelines
The Problem

Platform administrators spent hours authoring corporate course content, while low legacy test coverage led to continuous UI and workflow regressions.

The Architecture

Engineered production LLM course-generation pipelines with strict validation schemas. Overhauled the React learning portal and established rigorous Test-Driven Development (TDD) CI gates.

Verified Impact
• 20% administrator productivity increase
• Automated test coverage raised to 70%
• Full HubSpot CRM sync
03 / Technical Stance

Pragmatic engineering over resume-driven complexity.

I don't recommend Kubernetes clusters and 15 AWS microservices to an early-stage company. Here is the stack I use to deliver fast, reliable, and cost-effective software:

Primary Core

Ruby on Rails 8 + Hotwire

With Rails 8, the Solid Stack gives you background jobs, caching, and WebSockets without separate Redis clusters. Shipped with Kamal onto a $20/month VPS, it outperforms complex setups that cost $500/month.

Frontend & APIs

TypeScript, React, Node

When rich client-side state is strictly necessary, I build clean React and TypeScript interfaces backed by typed API endpoints, ensuring deterministic data contracts and zero hallucinated props.

AI Orchestration

MCP, Claude Code & Python

I build and integrate Model Context Protocol (MCP) servers, agent harnesses, and structured RAG pipelines. Human verification is strictly required at key boundaries to prevent model drift in production.

04 / Engagements

Clear scope. Zero hourly guesswork.

Diagnosis

48h Codebase Triage

For founders who just built an MVP with Cursor or Lovable and need an honest, deep technical inspection before launch.

$1,490 flat fee
  • • Static analysis & security scan
  • • N+1 & database index audit
  • • Multi-tenant isolation check
  • • 30-min Loom walkthrough
  • • 100% credited toward sprint
CORE ENGAGEMENT
Hardening & Launch

Production Sprint

Refactoring fragile code, setting up async workers, locking down auth, and deploying a battle-tested SaaS foundation.

$4,900 – $8,500 per sprint
  • • Full architecture refactor
  • • Solid Queue async workers
  • • Multi-tenant RBAC enforcement
  • • Automated TDD test suite
  • • Kamal / Docker zero-downtime deploy
  • • 14-day post-launch warranty
Leadership

Fractional Tech Lead

Senior engineering direction for post-launch startups that need a high-caliber technical partner without executive overhead.

$3,500 / month
  • • Weekly architecture triage
  • • Direct async Slack access
  • • High-complexity PR reviews
  • • MCP & AI pipeline setup
  • • Strictly capped at 2 founders
gx
Gabriel Dias
Tech Lead & Senior Full Stack Engineer
• Florianópolis, Brazil (Operating Globally)
• English (Fluent), Portuguese (Native), French

I am a Senior Full Stack Engineer and Tech Lead with 5+ years of experience steering distributed engineering teams (leading up to 9 engineers across 2 teams). I’ve taken monolithic Rails applications in production and navigated complex 11-month refactors with 99.9% continuous uptime, and restructured synchronous verification pipelines to run 12x faster (10 minutes down to 50 seconds).

I use AI agent workflows (MCP, Claude Code, multi-agent harnesses) daily to deliver code at 5x the speed of traditional agencies, but I pair it with strict human-in-the-loop validation, automated TDD test suites, and clean domain architecture.

When you work with GX Studio, you work directly with me. No junior developer handoffs.

Have a codebase that needs fixing?

Book a 30-minute discovery session directly on Google Calendar, or send an email with your repository details.

gxolivei@gmail.com