Speaking
Talks
Public talks at conferences and community events since 2012, in Spanish and English. Most of my speaking happens inside teams and with clients, as internal talks and training; the courses whose material is public are below. The sessions offered to organisers are on Sessionize.
Watch the recording Record
Public talks
The ones with a public record, most recent first. Each links only the material a source ties to it.
-
Albacete Power BI & Fabric User Group
Ingeniería del Software 2.0: diseñar, construir y validar productos con IA
Summary An adapted session on Software Engineering 2.0, exploring agentic development, Spec-Driven Development, engineering guardrails, and how AI is changing software delivery.
Original abstract
Llevamos años recortando en buenas prácticas de ingeniería para poder entregar a tiempo. Ahora que podemos delegar en la IA buena parte de la implementación, tenemos la oportunidad de revisar y la obligación de mejorar nuestros procesos de principio a fin.
Llevo más de veinte años desarrollando software y, en los últimos años, haciéndolo con IA, he tenido que replantearme muchas cosas: qué necesita mi proyecto para trabajar con agentes, qué necesitan los agentes para adaptarse al proyecto, cómo hacer que se respeten las decisiones técnicas y cómo validar que se entrega lo que se esperaba.
En esta charla compartiré lo que estoy poniendo en práctica y lo que sigo refinando. Hablaremos de desarrollo agéntico, Spec-Driven Development, guardarraíles duros y blandos y de la evolución del concepto de definición de un producto de software, con ejemplos reales y explicando las ventajas e inconvenientes de la IA en cada parte del proceso y qué sigue dependiendo completamente de nosotros.
La IA está cambiando la ingeniería del software y así es como algunos estamos diseñando cómo vamos a ejercer nuestra profesión en los próximos años.
-
IA contra IA: anatomía de un ataque dirigido a perfiles técnicos
Original abstract
Esta charla reconstruye una situación real en primera persona: una investigación nacida de una oportunidad profesional aparentemente legítima que escondía una trampa altamente sofisticada.
Related reading, on the same subject: Backdoor en Next.js y ataque a desarrolladores · hunting-the-hunter-episode-one
-
GitHub Copilot Dev Days Madrid
Pillado por una IA: detección y disección de un ataque dirigido a desarrolladores
Related reading, on the same subject: Backdoor en Next.js y ataque a desarrolladores · hunting-the-hunter-episode-one
-
Caught by an AI: Detecting and Hunting a Targeted Backdoor
Related reading, on the same subject: Backdoor en Next.js y ataque a desarrolladores · hunting-the-hunter-episode-one
-
Software Architektur im Stream
The Architecture of the Death Star - 20 Years of arc42 and What We Can Learn
Original abstract
In this session, we’ll examine the most iconic space fortress in film history through Juan’s complete arc42 documentation. This creative Star Wars project becomes an educational journey through the arc42 template - exploring how fictional architectures can teach us real lessons about software documentation.
-
UITP Global Public Transport Summit 2017
Summary Presentation of Mastria, the smart multimodal transport system developed at the Alstom Madrid Innovation Lab.
Teaching
Courses and workshops
Courses and talks given to teams, with their material in public repositories and articles.
- 01
GitHub Copilot Zero-to-Hero Crash Course
Given several times to engineering teams.
Practical workshop to go from zero to hero with GitHub Copilot. Focus: real engineering workflows, not just features.
- Orientation
- Foundations
- Prompting
- AI workflow quality
- Instructions
- Agents
- AI-ready repos
- MCP and orchestration
- Spec-driven development
- AI-SDLC
- 02
This project aims to document the fictional architecture of the Death Star by applying the arc42 model, a popular framework for structured software architecture documentation.
- All twelve arc42 sections
- English and Spanish
- 03
Running language models locally
Talks for teams, walking through the repositories.
VS Code–driven, terminal-first walkthrough of Docker Model Runner (DMR). One local runtime. Multiple clients. CPU or GPU. No cloud APIs.
- Docker Model Runner
- llama.cpp
- vLLM
Video courses in Spanish on YouTube: DADS, design patterns, Azure AI and Angular with .NET.
For organisers
Sessions I offer
Each one opens with the first lines of its abstract on Sessionize.
- 01
For years, we have been cutting back on good engineering practices to deliver on time. Now that we can delegate much of the implementation to AI, we have an opportunity to review, and an obligation to improve, our processes from end to end.
- 02
Issue to Merge: Agentic SDLC in Practice and What Breaks It
AI is changing how we build software, but the value only shows up when it supports the whole lifecycle and not only code generation.
- 03
Make Haste Slowly: Define Better to Build Faster
No product starts with a complete and final definition. Products are discovered, tested, and refined. The problem is not that decisions change. The problem is that they become scattered across roadmaps, tickets, documents, specifications, and conversations.
- 04
Engineering an AI-SDLC for Agentic Delivery
Reviewing a PR delivered by an AI agent is the wrong time to ask questions. What was it meant to change? Where did its context come from? How was it implemented?
- 05
The Architecture Before the Agent: Product Definition as Code
Agentic delivery has a context problem. A product decision must survive through architecture, data models, specifications, prompts, tests, security controls, and code.
- 06
AI-SDLC creates a traceability problem that becomes expensive as products evolve.
- 07
An agent can only be as safe as the engineering discipline it is given.