Based in MexicoAvailable for engagements

Senior Full-Stack & AI Engineer

Building AI-powered platforms, data infrastructure, and full-stack applications that run real business operations—not demos.

18+ Years Shipping Production Software

Top Rated | $100K+ | 6.9k+ Hrs on Upwork

Bilingual (English/Spanish)

Next.js 16 + React 19 + Python

View Featured Work Contact Me

Terraform/GCP Temporalio Workflows EDI X12 / AS2 Pinecone Vector DB MCP Servers Chrome Ext (V3) Playwright/Puppeteer Figma-to-Code Terraform/GCP Temporalio Workflows EDI X12 / AS2 Pinecone Vector DB MCP Servers Chrome Ext (V3) Playwright/Puppeteer Figma-to-Code Terraform/GCP Temporalio Workflows EDI X12 / AS2 Pinecone Vector DB MCP Servers Chrome Ext (V3) Playwright/Puppeteer Figma-to-Code Terraform/GCP Temporalio Workflows EDI X12 / AS2 Pinecone Vector DB MCP Servers Chrome Ext (V3) Playwright/Puppeteer Figma-to-Code

Evolution of Craft.

I am a software engineer focused on building systems that scale, from the UI down to the infrastructure layer. With over 18 years of experience (dating back to 2007), my career has evolved from managing high-volume agency deliveries to architecting distributed, cloud-native platforms.

How I Work: Async-friendly and spec-driven. Hand me a detailed brief and I deliver without hand-holding. I leverage modern AI tooling heavily—Claude Code and Cursor are core to my daily workflow (Opus for architecture, Sonnet for execution). I flag problems early, document handoffs, and write code the next engineer won't curse at.

Proven Web Vitals (Mobile)

Pacific NW Capital62 → 98 (LCP 1.6s)
KFB Law49 → 95 (TBT 30ms)
Estherian Clinic54 → 93 (LCP 2.7s)

Render-blocking JS, image pipelines, font loading, layout shifts — I diagnose it and fix it.

View Upwork Profile

2023 - Present

AI Engineering & Cloud-Native Architecture

Architecting 9-microservice distributed systems and integrating LLMs into production workflows. Building custom MCP servers, Temporalio durable workflows, and robust Python/FastAPI microservices deployed on GCP.

Model Context Protocol (MCP)Temporal & Vertex AIBigQuery & Cloud RunFastAPI & LangChain

2012 - 2025

Agency Anchor & Foundational Web

Anchored at a prominent US SEO agency for 13 years, building unusual range. Delivered BI dashboards, headless commerce, and internal AI tooling. Maintained production systems spanning WP, Shopify, and PHP over long lifecycles.

Long-term Production MaintenanceShopify & HeadlessInternal ToolingPerformance & Web Vitals

2007 - 2012

Early Career & Global Delivery

Began shipping web applications for international clients across the US, EU, and LATAM. Established a foundation in full-stack web development, localized payment integrations, and custom content management systems.

Bilingual Delivery (EN/ES)LAMP StackCustom CMSE-commerce

Technical Arsenal

Categorized by depth and current utilization in production environments.

Languages

TypeScriptPythonSQLHTML/CSSHCL (Terraform)PHP

Frontend

React 19Next.js 16Tailwind CSSshadcn/uiFramer MotionReactFlow

Backend

Node.jsFastAPIFlasktRPCExpressREST/SSE/WebSockets

AI & ML

Anthropic / OpenAI / GeminiMCP ServersLangChain / CrewAIPineconeHuggingFace

Databases

PostgreSQLBigQueryFirestoreDrizzle ORMPrismaRedis

DevOps & Cloud

GCP (Cloud Run, KMS, Pub/Sub)DockerTerraformTemporalioBullMQ

Featured Engineering

Deep dives into scalable data pipelines, automation, and AI integrations.

Data/MLWeb

3PL Business Intelligence Platform

High-performance data infrastructure tracking 2.26M transactions and $7.79M in revenue across 56 clients.

The ProblemA 3PL logistics operator needed to reliably sync and analyze 975K+ rows from Logiwa WMS in real-time without API timeouts.

The ApproachEngineered a Python ETL pipeline on Google Cloud Run feeding into BigQuery, surfaced via a Next.js App Router dashboard.

OutcomeDelivered a sub-200ms API response time on massive datasets, enabling real-time revenue tracking for 56 clients.

Role: Lead Full-Stack Engineer

Next.js 16React 19FlaskBigQueryCloud Run

Cloud/DevOpsAI/LLM

9-Microservice B2B AI Integration

AI-native B2B integration platform with visual pipeline builder and autonomous AI agents.

The ProblemA funded startup needed a reliable way to route AS2/EDI X12 protocols and connect live integrations with QVC, Tractor Supply, and ShipStation.

The ApproachArchitected a GCP-native distributed system using Temporalio workflows, six autonomous LangChain/CrewAI agents, and a ReactFlow visual builder with Monaco Editor.

OutcomeAchieved complete, observable workflow automation and seamless legacy EDI integration on modern infrastructure.

Role: Lead Systems Architect

FastAPITemporalioGCPReactFlowLangGraphEDI X12

AI/LLMWebData/ML

QueryChat AI

AI-powered data chat application with multi-DB abstraction and real-time LLM streaming.

The ProblemBusiness stakeholders lacked an intuitive way to query complex SQL databases without writing queries manually.

The ApproachBuilt a Next.js interface leveraging Anthropic's Claude to translate natural language into optimized SQL, executing securely via Drizzle abstraction.

OutcomeEnabled non-technical teams to generate live, accurate data reports in seconds via an intuitive chat interface.

Role: Full-Stack AI Engineer

Next.jsAI SDK (Anthropic)Drizzle ORMPostgresDocker

AI/LLMData/ML

Production MCP Server Suite

Cutting-edge AI-agent tooling bridging local LLMs with complex external APIs and enterprise data.

The ProblemLocal agents and LLMs lacked structured access to live business databases and vector stores.

The ApproachShipped four production MCP servers (Node.js/Python) allowing agents to query 100K+ client records, run AI-generated SQL, and pull data from a 352K-domain vector index.

OutcomeExpanded local AI agent capabilities, enabling autonomous data retrieval and structured reasoning.

Role: AI Tooling Developer

TypeScriptPythonModel Context Protocol (MCP)FastAPIPinecone

Web

Type-Safe Client Portal (OpenLedger)

Modern, end-to-end type-safe client management and financial portal.

The ProblemLegacy accounting workflows lacked modern developer experience and secure, real-time client collaboration features.

The ApproachDeveloped a strict type-safe monorepo integrating custom RBAC, responsive data tables, and AI-assisted data entry.

OutcomeDelivered a lightning-fast, WCAG AA compliant portal that increased client engagement.

Role: Lead Frontend Engineer

Next.jstRPCDrizzle ORMTailwind CSSPostgres

AutomationAI/LLM

AI Outreach Automation Engine

High-throughput automation pipelines for background processing and AI-synthesized scraping.

The ProblemManual lead generation and qualification workflows were bottlenecking sales operations and scaling poorly.

The ApproachImplemented a robust job queue system (BullMQ) driving stealth headless browser clusters to aggregate targeted data, synthesized via OpenAI.

OutcomeAutomated 85% of lead qualification, reliably processing thousands of background jobs daily.

Role: Backend & Automation Engineer

PrismaBullMQPuppeteerPlaywrightOpenAI / Anthropic

Additional Selected Work

SEO SaaS Tooling Suite

Dashboard and internal tooling for a prominent link-building agency.

ReactExpressPostgres

Visa Services Automation Platform

Bilingual, secure portal for managing immigration applications and documents.

Next.jsPrismaPostgres

Crypto Quant Trading Bot

Python trading system incorporating TA, textblob sentiment, and ML.

PythonccxtTransformers

Cross-Platform RTSP Dashboards

Ultra-low latency desktop media engineering for IP camera streams.

ElectronPythonOpenCV

Distributed Website Auditor

Turborepo architecture decoupling UI from intensive crawler worker nodes.

Next.jsRedisNode Workers

CBD Headless E-commerce

High-conversion headless build for a heavily regulated market.

ReactWooCommerce APIRedis

Ready to scale?

If you're building something real—AI systems, data platforms, complex B2B integrations, or need to rescue a project that's stuck—let's talk.

s@sgarza.com