I build companies with software and AI.
Entrepreneur.
Ten years building and scaling SaaS and AI companies from Barcelona: more than €10M in annual recurring revenue, a team of over a hundred people and customers in twenty-plus countries. I spend the rest of my time reading, listening and testing what comes next in AI, so the people around me do not have to.
- €10M+
- annual recurring revenue across the companies I run
- 5+
- SaaS and AI companies
- 100+
- people on the team
- 20+
- countries we sell in
- 60+
- countries I have visited
- 10+
- years building businesses
About
An engineer who ended up running companies.
I am a multidisciplinary engineer by training and a CEO by trade. For the last ten years I have built and scaled software companies: SaaS products, AI products and the teams behind them. Today that is more than five companies, over a hundred people and customers in more than twenty countries, run from Barcelona as CEO of Tendios.
I still write code, and I still read the papers. I spend an unreasonable amount of time on books, podcasts, videos and research about AI and technology, and I test everything I learn on real businesses with real revenue. That combination, operator plus student, is the whole point.
Before Barcelona I lived and worked in Vienna, Braunschweig and Coventry, and I have travelled to more than sixty countries. I speak Catalan, Spanish, English and German, and I studied at London Business School.
- Based in
- Barcelona
- Role
- CEO, Tendios
- Education
- London Business School
- Languages
- Catalan · Spanish · English · German
- Circles
- Círculo Ecuestre · Club Porsche
- Credentials
- PMP · SAFe Agilist · Certified ScrumMaster
Conversations
The questions I keep getting asked.
This is not a services page. But friends, founders and fellow members of the clubs I belong to keep asking me the same things, so here they are. If one of them sounds like your week, write to me.
What should AI actually change in my business?
Separate the demo from the deployment. Which processes a model can already run, which ones need a person in the loop, and what to leave alone for now.
How do I build a software product without a CTO?
Scope, stack, team and cost for a first version, plus the questions to ask any agency or developer before you sign anything.
Is this technology real or hype?
A plain-language read on tools, vendors and trends, from someone who has bought, built and thrown away plenty of them.
How do I take a company to twenty countries?
Product, pricing, payments, languages and the operational plumbing that turns a local business into an international one.
How do I use AI myself, today?
The concrete stack I use every day to write, research, code and decide faster, and how to set it up for a non-technical team.
Should I invest in, buy or start this?
An operator’s view on a deal, a product or an idea: what it takes to build, what it costs to run, and where the moat really is.
A WhatsApp message is the fastest way to start. Coffee in Barcelona works best.
Track record
Five-plus companies, one playbook.
I run a group of SaaS and AI companies, led from Barcelona and selling worldwide. Different products, the same principles.
- 01
Recurring revenue
Every company is a subscription business. Predictable revenue buys the right to think long term.
- 02
International by default
Twenty-plus countries taught us that languages, currencies and regulations are inputs, not a phase two.
- 03
AI in the operations, not the pitch
Support, content, analysis and code review already run with AI. It shows up in the margin, not in the deck.
- 04
Small teams, big leverage
A hundred people across five-plus companies means every team is small enough to move fast and senior enough to own it.
Thinking
What I believe about the next ten years.
I read, watch and listen to more about AI than is reasonable. This is what survives contact with running real companies.
- 01
Software stopped being the bottleneck.
Anything that can be specified can be generated. The scarce skills now are choosing the right problem, distributing the product and knowing when to stop.
- 02
Every company becomes an AI company, or a customer of one.
The question is no longer whether to use it. It is which of your processes it runs first, and who in your team is accountable for it.
- 03
Agents will do most of the work. People will do the deciding.
The founder of this decade designs the machine, sets its principles and makes the few decisions machines should not.
- 04
The advantage is adoption speed, not the model.
Models become commodities within months of release. The gap opens between companies that rewire how they work and those that wait for it to settle.
- 05
Small will beat big.
One person with the right system can run what used to take a building. Expect more companies, smaller and more profitable.
- 06
Europe is under-built.
One continent, many countries, under-served in almost every vertical. The next decade of software is a European opportunity.
Stack
The technology I work with.
Enough to build it, run it or judge it. Not a CV, a map of where I am comfortable.
- LLMs
- AI agents
- RAG
- Fine-tuning
- Evals
- Prompt engineering
- MCP
- Machine learning
- Deep learning
- Computer vision
- NLP
- Speech
- Vector databases
- AI governance
- Claude
- GPT
- Gemini
- Llama
- Mistral
- Anthropic API
- OpenAI API
- Hugging Face
- PyTorch
- TensorFlow
- scikit-learn
- LangChain
- LlamaIndex
- Claude Code
- Cursor
- TypeScript
- JavaScript
- React
- Next.js
- Node.js
- Vue
- Nuxt
- Python
- FastAPI
- Django
- NestJS
- Go
- Java
- Kotlin
- Swift
- .NET
- PHP
- Laravel
- React Native
- Flutter
- Tailwind CSS
- shadcn/ui
- GraphQL
- REST
- Prisma
- PostgreSQL
- MySQL
- MongoDB
- Redis
- Elasticsearch
- Kafka
- RabbitMQ
- Supabase
- Firebase
- AWS
- Google Cloud
- Azure
- DigitalOcean
- Vercel
- Cloudflare
- Docker
- Kubernetes
- Terraform
- GitHub Actions
- Playwright
- Vitest
- BigQuery
- Snowflake
- dbt
- Stripe
- HubSpot
- Salesforce
- Segment
- Figma
- Linear
- Notion
- Slack
- Google Workspace
- Zapier
- n8n
AI
LLMs · AI agents · RAG · Fine-tuning · Evals · Prompt engineering · MCP · Machine learning · Deep learning · Computer vision · NLP · Speech · Vector databases · AI governance
Models & tooling
Claude · GPT · Gemini · Llama · Mistral · Anthropic API · OpenAI API · Hugging Face · PyTorch · TensorFlow · scikit-learn · LangChain · LlamaIndex · Claude Code · Cursor
Build
TypeScript · JavaScript · React · Next.js · Node.js · Vue · Nuxt · Python · FastAPI · Django · NestJS · Go · Java · Kotlin · Swift · .NET · PHP · Laravel · React Native · Flutter · Tailwind CSS · shadcn/ui · GraphQL · REST · Prisma
Data & infrastructure
PostgreSQL · MySQL · MongoDB · Redis · Elasticsearch · Kafka · RabbitMQ · Supabase · Firebase · AWS · Google Cloud · Azure · DigitalOcean · Vercel · Cloudflare · Docker · Kubernetes · Terraform · GitHub Actions · Playwright · Vitest · BigQuery · Snowflake · dbt
Business systems
Stripe · HubSpot · Salesforce · Segment · Figma · Linear · Notion · Slack · Google Workspace · Zapier · n8n
Free-time project
Aurum VOS
One hundred companies at one million euros each.
In my spare time I am building Aurum VOS, a venture operating system: a shared platform, a foundry and governed AI agents that can launch, run and retire small software companies faster than people can. The goal is a portfolio of one hundred companies at one million euros of revenue each. It is my laboratory for everything I believe about AI, and it is public.
Visit aurumvos.com- 100
- companies, the target
- €1M
- revenue each
- Days
- from idea to a live product
Say hello
If you got here from a WhatsApp group, this is for you.
Friends, founders and fellow members of Círculo Ecuestre and Club Porsche: if AI or technology is on your mind, send me a WhatsApp. I am happy to think it through with you over a coffee in Barcelona. No pitch, no agenda.
WhatsApp +34 628 423 584
Or follow along with the 29,000 people who read my notes on LinkedIn.
