Available for remote work

Adrian
Wagemans

AI Application Engineer

LLMs · RAG · Agentic Workflows · Flutter · Automation

I build AI-powered products that actually ship. From LLM integrations and RAG systems to cross-platform mobile apps — I focus on solving real problems, not building demos.

View Projects LinkedIn →

What I've built

Real products. Real code. Real constraints.

CheeseTrace

Active development

Full-stack traceability and monitoring system for artisan cheese producers

Flutter FastAPI PostgreSQL BLE Tuya IoT OpenAI RAG pgvector
  • ▸ RAG layer on top of the in-process AI assistant — 1,445 chunks (27 Spanish DOP + 45 French AOP regulatory sheets + a full cheesemaking reference book) embedded with text-embedding-3-small, stored in pgvector with HNSW index; top 4 chunks (similarity > 0.55) injected as cited sources into the GPT-4o system prompt alongside the full batch context
  • ▸ BLE integration with Hanna HALO2 pH meter via Nordic UART Service — app detects stable readings (±0.02 over 5s) and syncs them to the backend automatically
  • ▸ Dual pH capture: BLE stream or photo of the meter display — GPT-4o Vision extracts pH and temperature from the image in <1s at ~$0.0001/call
  • ▸ IoT chamber monitoring via Tuya consumer sensors (~€15/unit): real-time webhook ingestion + 15-min async polling fallback, both normalized into a unified readings table
  • ▸ Generates a full batch PDF report per lot for regulatory traceability
Visit site → Private repo

TerraGuide

Active development

SaaS platform for rural municipalities to publish AI-generated audio guides

Flutter Node.js TypeScript FastAPI Next.js 14 Supabase OpenStreetMap
  • ▸ B2G SaaS — town councils create and publish multilingual audio tours with zero technical knowledge; onboarding a new municipality = 1 database record, zero code changes, instant white-label app with dynamic theming (primary color, logo, cover) applied at runtime
  • ▸ AI microservice (FastAPI + GPT-4o) generates full route narratives, per-point audio scripts and ES/EN/DE translations from a handful of admin hints — async job queue with atomic per-point persistence so partial results survive failures mid-generation
  • ▸ Flutter app (iOS + Android) with GPS auto-play: detects when the user enters a 50m radius of a route point and triggers the audio automatically; wakelock + background audio keep the session alive during the full walk
  • ▸ flutter_map + OpenStreetMap replacing Google Maps — zero cost, better rural coverage; polylines and route point markers rendered natively in Dart
  • ▸ Multi-tenant CMS (Next.js 14 App Router) with role-based access: superadmin manages all municipalities, each town council admin sees only their own data; Supabase Auth + JWT enforced at middleware and controller level
Visit site → Private repo

MyUrbania

Was in production · ~2 years

Cross-platform mobile application for GPS-guided urban audio tours

Flutter Firebase Firestore Cloud Functions Google Maps SDK TTS iOS Android
  • ▸ Architected and developed a cross-platform Flutter app delivering location-aware audio tours — shared codebase for iOS and Android with GPS-based navigation, distance calculations and automatic point progression as users move through a city
  • ▸ Data-driven content model: routes, points of interest, coordinates, media and metadata managed remotely in Firestore — new tours published without app updates
  • ▸ Built a TTS pipeline to generate narrated audio guides programmatically, enabling multilingual tour creation at near-zero marginal cost per city
  • ▸ Developed Firebase Cloud Functions for backend business logic and content processing; integrated Auth, Firestore, Cloud Storage and Analytics for scalable content delivery
  • ▸ Optimized for long walking sessions: async asset loading, client-side caching and efficient state management for smooth performance throughout the route
Visit site → Private repo

Koru — Skincare App & AI Analysis Pipeline

The Bubble Hub · 7 countries

Full Flutter skincare app with an integrated computer vision pipeline for skin analysis.

Flutter Computer Vision Roboflow Provider Machine Learning Python
  • ▸ Built the complete Flutter app from scratch: custom widgets, animations and multi-country release across 7 markets
  • ▸ Reached 200,000+ downloads across iOS and Android
  • ▸ Led the computer vision pipeline for skin analysis: data strategy, expert dermatologist labeling workflow and model training with Roboflow
  • ▸ Integrated the ML classification model into the mobile app for real-time skin analysis
Visit site →

Babymam — Pregnancy Tracking App

The Bubble Hub · Live

Android pregnancy app with 150,000+ downloads.

Android Java Mobile
  • ▸ Joined an ongoing production Android project and led UI improvements and new feature development
  • ▸ App reached 150,000+ downloads across multiple countries
Visit site →

WhatsApp AI Receptionist

MVP complete

24/7 AI receptionist for rural accommodation businesses via WhatsApp

n8n RAG OpenAI Supabase pgvector Next.js 15 WhatsApp API
  • ▸ RAG system with pgvector for property-specific knowledge retrieval (FAQs, pricing, availability)
  • ▸ Multi-tenant by design: adding a new client = 1 INSERT into the database, zero workflow changes
  • ▸ Handles lead capture, language detection (ES/EN/DE/FR) and owner handoff via WhatsApp commands
Private repo

How I build

The tools I reach for and why.

AI & LLMs

OpenAI API Claude API RAG pgvector LangChain concepts Agentic workflows n8n Claude Code

Mobile

Flutter Dart Riverpod GoRouter just_audio flutter_map

Backend

Node.js Python FastAPI Express TypeScript REST APIs

Data & Storage

Supabase PostgreSQL Firebase Firestore

Tools & Infra

Git Mapbox Google Maps API Railway Vercel Netlify

About

I'm an AI Application Engineer with a strong Flutter background, focused on building real products with foundation models.

I've shipped AI-powered mobile apps, led a computer vision pipeline for skin analysis with expert labeling, and built automation systems that handle real business workflows — including a multi-tenant WhatsApp AI receptionist and a SaaS platform for rural municipalities.

What sets me apart: I understand the business problem before touching the code. I've pivoted products when the market said so, designed multi-tenant architectures from scratch, and made build-vs-buy decisions — not just implemented features.

Based in Oviedo, Spain. Open to remote opportunities globally.

Experience

Apr 2024 – present

Founder & Developer

MyUrbania / TerraGuide

Jun 2022 – Dec 2025

Mobile Developer (Flutter)

The Bubble Hub

Dec 2021 – Feb 2022

Frontend Developer

TiendaCPU

Sep 2021 – Dec 2021

Freelance Frontend

Upwork