A working notebook of projects,
prototypes, and ideas.

Personal site · Ferdinand Straehuber

VecterAI

A working notebook of projects, prototypes, and ideas. I'm Ferdinand — IE University master's student in Technology & Global Affairs, building AI-native tools and vibe coding most of what's on this page.

Scroll
What This Is

A personal log of what I build,
what I break, and what I'm learning along the way.

This page is where I share my projects — finished, in-progress, and half-formed. Mostly AI-native applications: RAG systems, agentic workflows, predictive models, and the occasional weird side build that lives at the edge of what I'm studying.

I'm currently studying Technology & Global Affairs at IE University in Madrid, where I focus on how companies actually implement AI — the frameworks, governance, and structural decisions that make the difference between AI as a buzzword and AI as a working part of an organization. Most of the projects below sit at that intersection.

The stack I build with

Claude OpenAI Google Notion IE University
About

Student, builder,
and obsessive vibe coder.

I'm a master's student at IE University in Madrid, building things at the intersection of AI, governance, and whatever industries I happen to be curious about that week.

I'm currently in the Master in Technology & Global Affairs at IE University — a ten-month program that combines political science, international relations, economics, and law with the technologies reshaping how the world works: AI, fintech, cloud, and big data.

The program is built around three pillars — Citizenship & Rights, Geopolitics & Security, and Prosperity & Economy — with the governance and regulation of technology as the transversal thread through all of it. The program draws on input from the IE Center for the Governance of Change and collaborations with Google, Sciences Po, and the UN's International Telecommunication Union.

In practice, I focus on how companies actually implement AI: the frameworks, structural decisions, and governance choices that determine whether it works or quietly fails. Theory in the morning, prototypes in the afternoon.

Outside class I vibe-code obsessively. Most of what's on this site is the output of that — RAG systems, agentic workflows, hardware experiments, and full-stack tools I build end-to-end with Claude. Some of it is for clients (family business work, research collaborations), some is pure curiosity. All of it lives here.

01 — School

IE University, Madrid

Master in Technology & Global Affairs. Three pillars: Citizenship & Rights, Geopolitics & Security, Prosperity & Economy — with tech governance as the connecting thread.

02 — Focus

AI Governance & Implementation

How organizations actually adopt AI — the frameworks, regulation, and structural changes that decide whether it works or quietly fails.

03 — Method

Vibe Coding

Most projects start as a rough v0 written end-to-end with Claude. I iterate fast, ship rough, and learn from what breaks before designing anything permanent.

04 — Stack

RAG, Agents, Full-Stack

Tools I keep coming back to: Next.js, Supabase, n8n, the Claude API, OpenAI, Python for analysis, and whatever else the project needs.

Tools & school
Now

What I'm building
this month.

A snapshot of what's on my desk right now — active builds, research threads, and ideas that are still in the messy stage. Updated when things change. Finished and shipped work lives under Projects.

01
Active Build · Equine Tech

Three-sided stable management platform

The main thing I'm building right now. A full-stack platform connecting horse owners, trainers, and stable workers — built on a private RAG layer so all three sides see the same documents, contracts, and history. In-app chat, user profiles, and a bookable calendar for lessons, clinics, and farrier appointments. Currently wiring up the calendar engine and finishing the trainer-side onboarding.

RAG Next.js Supabase Chat Calendar
02
Research · Hardware

Polar-Equine — sensor capture & gait classification

A research data platform for sport horses. Connects to a Polar H10 Equine sensor over Web Bluetooth, captures heart-rate, R-R intervals, accelerometer, and raw ECG, then classifies gait patterns. Mobile-first PWA, v0 focused on "collect → clean → classify". Currently in Phase 1 of a 20-slice build plan. GitHub →

Next.js 14 FastAPI Web Bluetooth Supabase PWA
03
Client Work · Family Business

Straehuber AG — RAG for a German flooring company

Private RAG system for Straehuber AG, a German flooring business. Indexes every quote, supplier record, technical datasheet, and project history into a single searchable knowledge base — field teams can pull up any past project or material spec from their phone. Currently rolling out delta-sync so it stays live as new documents land.

RAG Private Deployment Mobile-first Document Indexing
04
Build · Construction

Construction spend dashboard + forecasting

A spend-analytics and forecasting dashboard for a construction company. Supplier invoices, category flow, rolling 2026/2027 forecasts, and a what-if simulator for price negotiation, maverick-spend formalization, and DPO tuning. Built on real project data (Ipanema, EDGE, Überseequartier). Live demo →

Spend Analytics Forecasting 3-way Matching What-if Sim
05
Internal Tool · Startup

StartupPanel — team & project tracking dashboard

An internal dashboard for tracking what my startup team is working on — sprint state, open blockers, who owns what, and live updates from connected tools. Designed for our team but built to scale to any small group. Still very much in the rough v0 stage.

Next.js Internal Tool Dashboard Team Management
Projects

Things I've built. The good,
the still-cooking,
and the abandoned.

A working list of personal and client projects — some shipped, some still in the messy stage, all built end-to-end. Mostly AI-native. Some heavily research-leaning. Status badges on each one.

Three-sided Equine Management Platform

A full-stack platform connecting horse owners, trainers, and stable workers. Built around a private RAG layer that ingests every document, training note, contract, and message — so all three sides see the same context. Includes user profiles, in-app chat, and a bookable calendar for lessons, farrier visits, and clinics. The core build I keep coming back to.

Stack Next.js, Supabase, Claude API, RAG, real-time chat, calendar engine
Status
In Development — main personal project. Calendar engine and trainer onboarding in progress.

Straehuber AG — Flooring Company RAG

A private RAG system built for Straehuber AG, a German flooring business. Indexes every quote, supplier record, technical datasheet, and past project into a single searchable knowledge base. Field teams can pull up any historical project or material spec from their phone before walking into a client meeting. Client work for the family business.

Stack RAG, Claude API, private deployment, mobile-first interface, delta sync
Status
Active — deployed and in use. Currently rolling out live document sync.

Polar-Equine — Sensor Data Collection & Gait Classification

A research-stage data platform for sport horses. Connects directly to a Polar H10 Equine sensor over Web Bluetooth to capture heart-rate, R-R intervals, accelerometer, and raw ECG data, then processes the signals and classifies gait patterns. Mobile-first PWA built on a "Collect → Clean → Classify" principle — v0 focuses on data acquisition, not health alerts. Codenamed La Fattoria / Sentavita.

Stack Next.js 14, TypeScript, Web Bluetooth, FastAPI (Python 3.11), Supabase Postgres, Vercel + Railway
Status
Research — Phase 1 of a 20-slice plan. View on GitHub →

Construction Spend Dashboard + Forecasting

A spend-analytics and forecasting dashboard for a construction company. Tracks supplier invoices, contract spend, category-level flow, and year-over-year trends with rolling 2026/2027 forecasts. Built on a real dataset of 10 projects (Ipanema, EDGE, Überseequartier) and 12 suppliers. Includes a "Recommended actions" engine and a what-if simulator for negotiating price, formalizing maverick spend, and tuning Days Payable Outstanding.

Stack Spend analytics, forecasting, PO-to-invoice 3-way matching, what-if simulator, Netlify-hosted
Status
Live demo — View dashboard →

StartupPanel — Team Tracking Dashboard

An internal dashboard for tracking what my startup team is working on — sprint state, open blockers, who owns what, and live updates pulled from connected tools. Built for our own team but designed to scale to any small group. Personal tool first, productized maybe later.

Stack Next.js, Supabase, TypeScript, integrations layer
Status
In Development — early v0. Functional but rough; iterating weekly.
How I Work

A few patterns across
most of what I build.

Most projects start the same way: a rough v0 written end-to-end with Claude, a private RAG layer at the foundation, and a real domain problem I actually want to solve. Then I iterate until something either works or earns the right to be shelved.

5
Active projects across personal, research, and client work
4
Built around a private RAG layer as the foundation
1
Research-stage hardware build (equine telemetry)
IE
Master's at IE University, Madrid — Technology & Global Affairs
Contact

Got a project in mind,
or just want to say hi?

Drop me a message — I'm always up for chatting about AI tooling, equine tech, vibe coding, the messy parts of building things, or anything in between. I read everything that lands in my inbox.

ferdinand.straehuber@gmail.com
IE University · Madrid, Spain
Open to collaborations, research, and conversations
Send a Message

I usually respond within a few days. No formality needed.