// Applied AI, no hype
I am Migue, a senior developer with 7+ years building software for companies. I document architecture, costs and failures from real products, and help teams apply AI without buying the hype.
Right now
Two products. Two problems that do not fit in a tutorial.
Work with me
I work directly with companies on two fronts: I build automations and prepare teams to use them with sound judgment.
“You answer the same question twenty times a day. You copy data from an email into a spreadsheet. You chase invoices, reminders and reports a machine could handle alone.”
I review your operation, prioritize the process with the clearest return, and build the automation around the tools your team already uses.
Published prices · Limited capacity
See services and pricingWe start with a real company process, work through it in a practical workshop, and leave behind a playbook the team can repeat without depending on me.
Remote · ES/EN · Tailored program
Book a diagnostic≈ COP 1,000,000
I review your operation, find the 3-5 tasks eating most of your hours, and deliver a prioritized plan with estimated hours saved and tooling costs.
≈ COP 2 – 6 million
I build the automation end to end with n8n, AI agents and your current tools. You get documentation and a short how-to video for your team.
≈ COP 1.2 – 3 million
I maintain, monitor and extend your automations every month. When your tools change, I adjust things before you notice.
I take at most 3 retainers at a time, so I can respond fast and with quality. This is not an agency: you work directly with me.
Recent work
These are not Dribbble concepts or conference demos. They are systems with users, billing, migrations and failures that need fixing on a Monday morning.
Case 01 · B2B SaaS
Live product
A platform that turns URLs and documents into AI avatar videos, GEO packages and complete publish-ready pages.
Visit IntelligentVideos.aiMy work
I build and maintain the web product and backend: authentication, credit billing, generation queues, Claude, Tavus, Bunny Stream and Stripe integrations, plus cost observability.
What it already solves
Video Creator, GEO Optimizer and Page Transformer share one account, one credit system and an asynchronous flow that delivers video, code and content.
Stack and systems
Case 02 · Health marketplace
Mobile app and migration in progress
A bilingual platform connecting patients with recovery homes and services for medical travel.
Visit Recovery Care SolutionsMy work
I built the React Native client and am now replacing the legacy backend with a TypeScript API using onion architecture, role-based authorization, SQL migrations and tests against the previous system.
What it already solves
The web platform already supports stay search and bookings. The mobile app is in pre-launch, while the new API advances route by route without breaking current flows.
Stack and systems
Parking management with an Android operator app, web dashboard and paying customers.
Own product · Astro · React Native · Node.js
Visit ParqandoMobile-first habit tracking with leagues, streaks, statistics and rankings.
Own product · Next.js · Prisma · PostgreSQL
Open Warrior TrackerWhat I'm building and studying right now: production-grade systems, not just POCs.
Building production-grade serverless architectures on AWS and exploring multi-tenant SaaS patterns with Lambda, API Gateway, and RDS Aurora. Studying cost optimization strategies for bootstrapped startups. Target: AWS SA Professional certification (2026).
Stack: Next.js frontend, FastAPI backend, AWS Lambda, PostgreSQL. Implementing semantic search over internal documentation and experimenting with prompt engineering and retrieval optimization.
Integrating OpenAI and Anthropic APIs into existing TypeScript/Node.js services. Building reusable AI feature modules for SaaS applications with focus on context management, token optimization, and error handling at scale.
Learning AWS SageMaker, model deployment patterns, and monitoring. Exploring fine-tuning workflows with Hugging Face and LoRA/PEFT. Target: AWS Machine Learning Specialty certification (2026).
Designing cost-effective AI-native SaaS architectures. Comparing serverless vs. containerized inference workloads and building deployment pipelines for LLM-powered features.
Work log
I write when I have something to demonstrate, not to fill an editorial calendar.
March 2, 2026 · 4 min read
March 2, 2026 · 2 min read
The person behind the code
I have spent 7+ years building software that lives in production: TypeScript, React, Node.js and AWS across retail, IoT, telecom and SaaS.
Today I work on AI products with the uncomfortable parts included: authentication, billing, queues, observability, support and users who do not read the manual. The demo is usually the easy part.
I also teach applied AI every week at a programming academy and build my own products in public. That is where everything I share comes from: real decisions, real costs and failures that left a scar.
Does your company need to build, automate or prepare the team to use AI with sound judgment? Let us discuss the problem before discussing tools. And if what you need is a senior remote developer (cloud + AI) on your team, write to me too.
Contact
You do not need to arrive with the solution defined. Describe the process, bottleneck or what the team needs to learn.