# Javier Lougedo

> Software engineer with an MBA. I ended up spending most of my time where products make money, and the rest as a professor explaining AI to lawyers and business people.

Madrid, Spain · remote-first, EU hours. Links: [LinkedIn](https://www.linkedin.com/in/javierlougedo/) · [GitHub](https://github.com/Lougedo) · [Email](mailto:javier.lougedo.lorente@gmail.com) · [Email for students](mailto:javieralejandro.lougedo@unir.net) · [CV (PDF)](https://javierlougedo.com/cv/javier-lougedo-cv.pdf).

## Now

- Senior SDE & Tech Lead at Typeform · Blocks
- Professor of applied AI at UNIR
- Building unsaid · a daily game with AI

## Experience

- **Senior SDE & Tech Lead · Blocks · Typeform** (2025–present)
  - Tech Lead for Blocks, the shared platform behind the form builder and renderer; drive cross-repository delivery across React/TypeScript frontends, Forms API, submissions and Go services.
  - Built the foundation for Stripe subscription payments end to end: plan selection, Stripe Elements, 3DS flows, server-side creation and verification, and a provider abstraction that keeps PayPal viable without a rewrite.
  - Led PayPal integration work across builder and renderer: connection management, safe disconnection, live-preview behaviour, SDK error handling and payment-service monitoring.
  - Technical bridge across Product, Design, Data, SRE and Integrations: shaping scope, estimating work, reviewing PRs, mentoring teammates and unblocking delivery.
- **SDE · Site team · Typeform** (2024–2025)
  - Shipped growth experiments end to end on the public site with Optimizely (hypothesis, implementation, targeting, tracking validation, QA, rollout and the ship/kill call) across signup, navigation and homepage surfaces.
- **Professor of data analytics and applied AI · UNIR** (2024–present)
  - Teach applied AI, data analytics and programming fundamentals in UNIR's Master's Degree in Business Innovation and Transformation with Artificial Intelligence, to cohorts of ~110 students from business, economics and law backgrounds.
  - Build retrieval and agentic workflows with students: chunking, embeddings, vector stores, retriever agents, guardrails and MCP, with n8n, Flowise, Make and locally-run models.
  - Built a multi-agent assessment pipeline that drafts individual rubric-based feedback for the full cohort; supervise Master's theses on applied AI and business.
- **Freelance developer · MIKE** (2024–2025)
  - Platform giving real estate workers instant AI-powered legal consultation.
- **Web analyst & developer · Openbank** (2023–2024)
  - Launched and led the cards vertical for one of the group's banks.
  - Standardised imports across all web teams and documented the frontend codebase.
- **Frontend developer · HeyTrade** (2021–2023)
  - Built the mobile and web trading apps with React and React Native; supervised trainees and ran tech interviews.
  - Led the credit/debit card payments integration with checkout.com.
  - Improved onboarding conversion by over 50% through user-data analysis and ML, and automated monthly investor reporting with Python and Pandas.

## Education

- **Tech MBA · Universidad Internacional de La Rioja (UNIR)** (2025–2026): Final thesis on Intelligence-as-a-Service architectures built on MCP.
- **Computer Engineering · Universidad Autónoma de Madrid** (2017–2021): Additional summer coursework in Data Science, Machine Learning and Cloud.

## Skills

- Programming languages: TypeScript, JavaScript, Node.js, Go, Python, SQL, HTML/CSS
- Frameworks and platforms: React, React Native, Next.js, AWS, Kubernetes
- Growth and payments: A/B testing, Optimizely, LaunchDarkly, Stripe, PayPal, Subscription billing
- AI and automation: n8n, Flowise, MCP, Ollama, ChromaDB, Claude Code

## Projects

- [CNN and people-counting notebooks](https://github.com/Lougedo/cnn_samples_colab) (2026, UNIR · IASP): Two Colab notebooks for students who do not code. Design and train a CNN on chest X-rays (PneumoniaMNIST) by moving form controls, and count people in a video with YOLO by zone and at a line, exporting counts only. Generated by script and verified with zero errors after two adversarial reviews.
- [Teaching RAGs with n8n](https://github.com/Lougedo/rag_nexora_industrial) (2026, UNIR · IASP): Two RAGs built in n8n and Qdrant for class. Clinical guidelines (984 documents, 128,787 chunks, interchangeable drafting models) and industrial compliance (100 documents, three answer profiles, 33 of 35 on the evaluation). Both corpora carry planted traps, so students watch a RAG fail.
- Class prep with Claude Code (2026, UNIR, In use): A Claude Code workflow for preparing each class. Nine fixed rules (a theory ceiling, one tangible deliverable, 120 minutes that add up, never invented data), four phases I sign off one by one and a nine-check quality gate. It builds the deck with speaker notes (python-pptx) and generates and verifies the notebooks by script.
- Subscription payments with Stripe (2025, Typeform · Blocks): The foundation for Stripe subscriptions in Typeform forms, end to end. Plan selection, Stripe Elements, 3DS, server-side creation and verification, behind a provider abstraction that keeps PayPal viable without a rewrite.
- Multi-agent feedback pipeline (2025, UNIR): Reviews every submission against the module's rubric and drafts individual, criterion-by-criterion feedback for a cohort of ~110 students. I review and sign off every grade; the system removes the bottleneck.
- MIKE, an AI legal advisor (2024, Freelance): Legal-advice SaaS for real estate professionals, built freelance between November 2024 and February 2025.

## Teaching

I teach applied AI, data analytics and programming fundamentals to non-technical profiles in UNIR's Master's Degree in Business Innovation and Transformation with Artificial Intelligence. Upcoming sessions and the material I prepare live here.

- [IASP · Applied AI in Productive Sectors](https://javierlougedo.com/university/iasp/) (7 sessions of 2 hours): AI in six sectors: banking and insurance, retail, education, health, industry and agriculture, and transport and logistics. Every session pairs a sector with a no-code practice (Colab, Make, n8n, KNIME, a RAG) and ends with something that works.
- [TAAVD · Advanced Data Analysis and Visualisation Techniques](https://javierlougedo.com/university/taavd/) (7 sessions of 2 hours): Data analysis and visualisation for business: from the types of analysis and KPIs to dashboards, with Colab and Python, Power BI, Looker and KNIME, and storytelling to tell what the data says.
- [FPA · Programming and Algorithm Fundamentals](https://javierlougedo.com/university/fpa/) (5 sessions of 1 hour): Algorithms for non-technical profiles: what an algorithm is, what it costs in time and memory (big-O, intuitively), from the machine to Python and AI, data structures and strategies. One Colab notebook per session.

Upcoming sessions:

- Wed, 30 Sept 2026, 19:00 CEST · FPA · S1 · Qué es un algoritmo
- Wed, 7 Oct 2026, 19:00 CEST · FPA · S2 · La pirámide: de la máquina a Python y la IA
- Wed, 14 Oct 2026, 19:00 CEST · FPA · S3 · Estructuras de datos
- Wed, 21 Oct 2026, 19:00 CEST · FPA · S4 · Complejidad: notación O, buscar y ordenar
- Wed, 28 Oct 2026, 19:00 CET · FPA · S5 · Estrategias: fuerza bruta, divide y vencerás, voraz y programación dinámica

Materials (public links), by subject and session:

- FPA · S1
  - Adivina el número: una a una frente a por la mitad · Notebook: https://colab.research.google.com/drive/1pUul1XtkN4I7MigOLy26ixV6CoBh0yjf
- IASP · S7
  - Una CNN por dentro, con radiografías · Notebook: https://colab.research.google.com/github/Lougedo/cnn_samples_colab/blob/main/notebooks/01_laboratorio_cnn_radiografias.ipynb
  - De una sentada frente a paso a paso: pedirle el trabajo entero a la IA · Guide: https://drive.google.com/file/d/1GG8BUBJ43Bbhr0j4-3ysiLYFqWKK2Qkm/view
  - Contar personas en un vídeo · Notebook: https://colab.research.google.com/github/Lougedo/cnn_samples_colab/blob/main/notebooks/02_conteo_personas_cafeteria.ipynb
  - Demo multiagente: cinco revisores critican la solución · Guide: https://drive.google.com/file/d/1dwlThEkTjjEhefz70jVK8u3hTMdvchpH/view
  - Hoja del alumno: los dos cuadernos en una página · Guide: https://drive.google.com/file/d/1C-JVF1T_yHkJNRFprrQsLiu3RT-ScyLp/view
  - Método: resolver las actividades con IA (design thinking, siete pasos y prompts) · Guide: https://drive.google.com/file/d/1vYc5zWpMdq6dDNRKQ5YXccji7k31svgy/view
  - Las actividades, paso a paso, y una CNN por dentro · Slides: https://drive.google.com/file/d/1g_UjuPkDVxqd9LUO6CQ2lgRt5dCT-LxE/view
- IASP · S6
  - Make: la torre de control de envíos · Guide: https://drive.google.com/file/d/1oVEZjeB2u1VjSzgUutjxLDw3QdSVsrhl/view
  - IA en Transporte y Logística, y cierre de la asignatura · Slides: https://drive.google.com/file/d/1YnuysTi1rxm8XMUZzdncRuQPWk03hs66/view
  - Prompts para que Claude monte la torre de control en Make · Guide: https://drive.google.com/file/d/1TWX3kr19HJL4L41SJYUz_HUV_2mQNGIe/view
  - Escenario Make: segunda vuelta con tres jueces, responsable y registro · Workflow: https://drive.google.com/file/d/1AvV0rnVpjWot1eGztukIEHUCicE8emMA/view
  - Escenario Make: torre de control de envíos (con plantilla CSV) · Workflow: https://drive.google.com/file/d/17r8FRCsvUIJyT1HPJBD1kZK4a-3BpH6i/view
- IASP · S5
  - RAG de cumplimiento en n8n: montaje, cinco preguntas, trampas y evals · Guide: https://drive.google.com/file/d/12KeyLbMko2Xbh69bk1N6wBkinHjL8efP/view
  - IA en Industria y Agricultura, y un RAG de cumplimiento · Slides: https://drive.google.com/file/d/1S_0SyLZu-GSeVEc5AdwL-LuJRK6pjxGG/view
  - Flujo n8n: RAG de cumplimiento industrial · Workflow: https://github.com/Lougedo/rag_nexora_industrial/blob/main/n8n/IASP_S5_RAG_cumplimiento.json
  - RAG de cumplimiento NEXORA Industrial: 100 documentos, montaje y evaluación · Workflow: https://github.com/Lougedo/rag_nexora_industrial
- IASP · S4
  - RAG paso a paso en n8n: dos guías de diabetes y cinco experimentos · Guide: https://drive.google.com/file/d/1sHQrFsOCQ7-OHf77eLMu4XmQRGr9qorp/view
  - Monta tu RAG: guía completa (n8n + Qdrant, tus documentos y costes) · Guide: https://drive.google.com/file/d/1Xt1QELKiYXvgrf0ykgRm3Yl7Hjj19Yj7/view
  - IA en Salud y un RAG en n8n · Slides: https://drive.google.com/file/d/12mnHeI0lqAHB_K7g6r_f1eNRxikd5K60/view
  - Flujo n8n: RAG sobre guías clínicas (práctico de clase) · Workflow: https://github.com/Lougedo/rag_nexora_salud/blob/main/n8n/IASP_S4_RAG_guias_clinicas.json
  - RAG multimotor NEXORA Salud: 984 documentos, flujos n8n y Flowise · Workflow: https://github.com/Lougedo/rag_nexora_salud
- IASP · S3
  - Fuentes: el ensayo de Amodei y las reacciones · Guide: https://drive.google.com/file/d/189gVutqeuoCeAlVogSFeTfnTVXAfw0Ji/view
  - Fuentes para el cuaderno de Educación (Gemini Notebook) · Guide: https://drive.google.com/file/d/1miIhZj51edpDthX3adzcizeL5pLKJzcv/view
  - Make: feedback automático de entregas (diseño, rúbrica y prompt) · Guide: https://drive.google.com/file/d/10doyXBFNycpc7k9GZolA8vcf1kYmc3Bf/view
  - Educación, Gemini Notebook y RAG · Slides: https://drive.google.com/file/d/1uFmfshFp8x-GWwITrFKGwvtNfYm9YS4L/view
- IASP · S2
  - Make paso a paso: del correo a la hoja, y la IA que clasifica · Guide: https://drive.google.com/file/d/1ZKLnbt8wPB3BNjodU3PrATbKaYUO7smu/view
  - Banca (cierre) y Retail · Slides: https://drive.google.com/file/d/1c9YtC8cDOOXaT60iA5EOdhR3wClalg1R/view
- IASP · S1
  - IA en Banca y Seguros · Slides: https://drive.google.com/file/d/1O3-NVbGz6PM2a8R6_PiWI62fdM1AOgqd/view
  - ¿Contratará este cliente? Scoring bancario, fuga de datos y explicabilidad · Notebook: https://colab.research.google.com/drive/1EI0fAXfOkzY360-ELRYbK8n1fDCVKg_z
- TAAVD
  - Visualizaciones avanzadas con Python: Superstore · Notebook: https://colab.research.google.com/drive/1i6WKoTZdWlWaAWZ7CgYBt5a5XwAcrLkk
  - Segmentación de clientes de un centro comercial (K-Means) · Notebook: https://colab.research.google.com/drive/1YIAofEYn3ZwcEyRVr7kpcgrR3H_XN-uD
- TAAVD · S1
  - Análisis descriptivo en profundidad: Superstore · Notebook: https://colab.research.google.com/drive/1afGZENkdjbCzR3V9syw3_hEOumAdk7EH
- TAAVD · S5
  - Churn en telecomunicaciones: cinco modelos y exportación a KNIME y Power BI · Notebook: https://colab.research.google.com/drive/1PPf4ExfRJxJqDP4YuxRfp3LRoXmjuNYu

## Lab

- [Design system](https://javierlougedo.com/lab/design-system/) (In use): The tokens and components this site is built with
- [Ask my site](https://javierlougedo.com/lab/ask/) (In progress): Questions about my work, answered from this site's facts
- unsaid (In progress): A daily game with AI
- PDF tools (In progress): Utilities for working with PDF documents

## Pages

- [Home](https://javierlougedo.com/) · [Spanish](https://javierlougedo.com/es/)
- [CV](https://javierlougedo.com/cv/) · [Spanish](https://javierlougedo.com/es/cv/)
- [University](https://javierlougedo.com/university/) · [Spanish](https://javierlougedo.com/es/university/)
- [University · IASP](https://javierlougedo.com/university/iasp/) · [Spanish](https://javierlougedo.com/es/university/iasp/)
- [University · TAAVD](https://javierlougedo.com/university/taavd/) · [Spanish](https://javierlougedo.com/es/university/taavd/)
- [University · FPA](https://javierlougedo.com/university/fpa/) · [Spanish](https://javierlougedo.com/es/university/fpa/)
- [Lab](https://javierlougedo.com/lab/) · [Spanish](https://javierlougedo.com/es/lab/)
- [Ask my site](https://javierlougedo.com/lab/ask/) · [Spanish](https://javierlougedo.com/es/lab/ask/)
- [Design system](https://javierlougedo.com/lab/design-system/) · [Spanish](https://javierlougedo.com/es/lab/design-system/)
