IaaS · PaaS · SaaS — the full cloud stack
We design and run the whole stack around these three layers — IaaS foundations on AWS with Terraform and serverless infrastructure, PaaS platforms built on event-driven microservices and CI/CD, and SaaS products shipped end to end — like QuizHub, RedLie, and Cargonaut below.
The Prompt Engineering learning path — a hands-on, seven-part series. Describe what you want in plain English and let Claude Code build it: a web app, online, with a backend (Node.js, Python, or containers), automated with CI/CD, and driven like a pro. Read each as an article, or run it as a slide deck.
Start here. Describe what you want in plain English and let Claude Code build a Next.js app — install, prompt, preview, iterate. Covers npm/package.json and project structure. Ends with the app running locally.
Beginner · read or presentThe base layer under every app: VPC, subnets, security groups, S3, EventBridge, IAM, RDS. Which pieces to build with Terraform (shared, stateful, long-lived) vs CloudFormation/SAM (per-service compute) — and how services plug in through SSM.
Infrastructure · read or presentNow get it online. What Vercel and Supabase actually do, how to deploy, and the AWS services (S3+CloudFront, Amplify, Lambda, ECS Fargate, DynamoDB, Cognito) that do the same jobs.
Foundations · read or presentAdd a backend in the same codebase. Build a Notes API with Route Handlers — a swappable data layer, dynamic routes, validation. The fullstack-JavaScript path.
Node.js API · read or presentA standalone serverless service, built the way JunctionNet ships APIs: layered handlers → services → repositories → models with a middleware factory for DI. API Gateway + Lambda (Python) + DynamoDB via SAM.
Python API · read or presentFor a real-server app (NestJS, Express, FastAPI, workers, Redis, Kafka): containerize with Docker and run on ECS Fargate behind a load balancer, deployed from CI. Modeled on jnet-transport-backend.
Docker + ECS · read or presentLet services react to each other without calling each other. The event-driven pattern on Amazon EventBridge — naming source and detail-type, when and where to publish, extending a system by adding consumers, and forwarding events to OpenSearch for analytics.
Event-Driven · read or presentThe capstone: level up how you drive the tool. Permission modes, ultrathink, auto mode, remote sessions, ultracode multi-agent workflows, the /run → /verify → review loop, and the anatomy of a CLAUDE.md.
Power user · read or presentAutomate it. One GitHub Actions pipeline — CI checks on every PR, preview environments, keyless OIDC auth to AWS, protected production, and rollable releases — wired to each stack from Parts 1–6.
Automation · read or presentProve who is calling, the right way for each caller: Bearer JWTs for services, cookies for browsers, static API keys for third parties, HMAC signatures for webhooks — and why the Authorization header almost always wins. Ties into the platform Lambda Authorizer.
Security · read or presentWe're a software engineering team specializing in cloud-native platforms for the logistics and freight-forwarding industry. We've designed and delivered full-stack products from blank repo to production — handling everything from AWS infrastructure and serverless microservices to polished customer-facing portals.
Our stack is Python + AWS Lambda on the backend, Next.js / Vue.js 3 on the frontend, and Terraform + Terragrunt for infrastructure. We have deep experience with event-driven architectures (EventBridge), CI/CD automation (GitHub Actions), and integrating AI capabilities (Anthropic Claude API).
We enjoy working with early-stage companies and product teams that need a partner who can think at the architecture level and write the code. Production-tested technologies across the full stack — cloud to frontend.
Technical writing on software architecture, API design, cloud infrastructure, and engineering practices.







































The products we design, build, and run — architecture, backend, frontend, and DevOps.
A freight operations OS for Mexico-bound shipments. Full-stack product with 6 Python Lambda microservices, an AI-powered Next.js app (React 19), a customer portal, and a Terraform-managed AWS platform. Pilot customer: SANHUA.
A large-scale logistics platform monorepo with ~18 Python Lambda microservices, 5+ Vue.js 3 portals (EntryWriter, Organizer, Pricing, UDM, Workspace), and a custom CI/CD framework.
Free software to compute the índice de disponibilidad léxica (lexical-availability index). Upload a CSV, compute the indices, and export the results to Excel.
Kahoot-style live multiplayer trivia. A host goes live and shows a QR code; players join from their phones — no login — answer timed questions, and a podium appears at the end.
Personal habit and finance tracker — log daily habits and spending, with all state persisted to a Google Sheet and visualized with charts.