Consulting · Medellín, Colombia
Architecture & engineering for AI and data platforms
Based in Medellín, Colombia, I help teams across LATAM and other time zones design and build AI platforms, data pipelines and ML systems. I can lead architecture decisions with your team or build a platform through deployment.
Goals and constraints.
An agreed scope: what the team needs, which data exists and what constrains the solution.
Architecture and decisions.
Architecture, alternatives and a phased plan with verifiable criteria.
Software, tests and operation.
Software integration, tests and documentation so the team can operate and maintain it.
Types of project
End-to-end AI platforms
Applications with LLMs, RAG and agents. We define test data, evaluation criteria, usage limits and the product that connects the model to its users.
Data infrastructure
Ingestion and transformation pipelines, data validation, APIs, and batch or streaming execution according to the product’s needs.
ML systems & MLOps
Reproducible training, deployment, monitoring and retraining criteria. We define what to measure and how to detect failures once the model is in use.
Product & platform architecture
Review of services, APIs, data contracts and plugin systems. I document alternatives, costs and limits so the team can decide and maintain the system.
Applied optimization
Hyperparameter tuning, feature selection and shift planning. I translate business constraints and objectives into models that can be evaluated.
AI strategy & design
Prioritizing use cases, defining risks and testing with success criteria before committing to a full project.
How I work
01
Define the problem: existing systems, available data and constraints. Deliverable: an architecture proposal and work plan.
02
Build in phases: components your team can test, model evaluation and operational controls. Each phase has a reviewable result.
03
Prepare operations: documentation, design decisions, risk review and handoff to the team that will maintain the system.
Published work
These projects let you inspect my code, design decisions and usage examples.
A visual pipeline builder in alpha, with standalone Python export.
Evolutionary hyperparameter optimization and feature selection for scikit-learn.
Staffing, shift planning and resource allocation under constraints.
Frequently asked questions
Do not see your question here? Reach out and we will talk through your specific case.
What kind of projects do you take on?
+
AI platforms, data infrastructure, production ML systems, complete software products, and complex internal tools. Also architecture leadership or review. I do not take on one-off support work, small automations, or simple website builds.
Do you design the architecture or also build the system?
+
Both. I can lead architecture decisions alongside your team, or design and build the complete platform end to end: data, models, services, and product. Most projects combine the two.
What evidence is there that you can build systems like this?
+
My libraries and Ciaren have public code and documentation. This page links to their examples and design decisions. The experience section details my work at EPAM, Globant, Rappi and Avianca.
Do you work with companies outside Colombia?
+
Yes. I am based in Medellín, Colombia. I work with teams across LATAM and other time zones; we agree on hours and working sessions according to the project’s needs.
How does a typical project start?
+
With a conversation about the business goal and the current technical state. From there comes an architecture definition: assets, constraints, risks, and a concrete design; and a phased construction plan with explicit production criteria.
Are you building an AI or data platform?
Tell me what the team needs, which system exists and its main constraint. We can use that to define the scope of our work together.