Available for selected projects
I’m Edu García, a Full Stack Developer based in Seville. I turn manual processes and unclear digital products into fast, measurable and maintainable systems.
Projects
I do not show empty mockups. I show systems, libraries and tools designed to solve concrete problems: invoicing, data, conversion, automation and daily operations.
Open-source library · Python · Verifactu
MIT library to generate Verifactu invoicing records inside your own software: chained SHA-256 fingerprint, tax QR, XML, XSD validation and optional sending/signing modules. Built for developers; it is not positioned as a closed SIF or invoicing SaaS.
FastAPI · eCommerce · automation
Private abandoned-cart recovery and product-recommendation engine for regulated eCommerce. Includes eligibility logic, cart deduplication, shipping-condition calculation and transactional communication preparation.
Desktop · FastAPI + pywebview
Desktop invoicing app for a local business: service catalog, backups, document generation, adaptable fiscal logic and native installer packaging.
ETL · dashboards · reporting
Business intelligence platform for internal operations: data extraction, normalization, sales metrics, customer metrics, warehouse analysis and executive monitoring with charts and maps.
Next.js · digital forensics
Tool to analyze signals of manipulated content: metadata, visual patterns, technical consistency and behavioral signals. Designed to explain the result, not only provide an opaque score.
PWA · image analysis
Nutrition-tracking app with daily logging, weight trend and AI-assisted image analysis to recognize meals from photos.
AI agents · RAG · workflows · intelligent backends
A work line focused on integrating artificial intelligence into real processes: internal assistants, ticket classification, document data extraction, controlled response generation, semantic search over private documentation and automations connected to custom APIs.
Featured project
Open-source (MIT) Python library to generate Verifactu invoicing records —chained hash, tax QR, XML and signature— without sending your invoices to a third party or paying a per-tax-ID fee. Built for developers who integrate, not as a closed SIF.
Honest status: core validated with official AEAT material and submission validated live against the real test environment (invoices accepted with official CSV). Production use depends on each developer's certificate and integration.
Services
I work especially well where backend, data and business meet: manual processes, slow systems, broken integrations, poor reporting or fiscal software that is hard to implement.
Architecture, authentication, endpoints, data models and business logic with FastAPI or Laravel.
Workflows that save hours: classification, extraction, alerts, document generation and repetitive tasks.
ETL, dashboards and clean metrics so operations can make decisions with reliable data.
Chained fingerprints, tax QR, XML, validations and technical criteria for fiscal integrations.
Cart recovery, business rules, recommendations, measurement and conversion improvements.
Desktop tools, private panels and custom utilities when generic SaaS falls short.
Applied AI
I do not use artificial intelligence as decoration. I integrate it into software, processes and data to reduce manual work, improve answers, extract information, connect systems and build internal tools with measurable value.
Agents connected to business rules, APIs and databases to execute tasks, query information and support internal teams without losing control.
Semantic search and answer systems over PDFs, manuals, internal policies, catalogues, FAQs or technical documentation.
Extraction, classification and summarization of invoices, emails, tickets, orders, contracts and operational documents.
Support assistants, recommendations, product content, segmentation, cart recovery, query analysis and post-sale workflows.
Image analysis to classify content, detect patterns, recognize products or support applications with visual data.
Language model integration, controlled prompts, validations, logs, costs, limits, security and deployment inside existing backends.
Method
Before writing code, I organize the problem. Then I build a usable, measurable first version that can grow without turning into technical debt.
Understand the process, users, data, current errors and real cost of the problem.
Define entities, integrations, boundaries, risks and the first viable version.
Iterative development with tests, real data and documented decisions.
Deployment, documentation, basic training and the next technical roadmap.
Profile
My advantage is not just coding. It is understanding how real processes fail: orders, data, customers, reporting, warehouses, emails, incidents and business decisions.
I work with a practical mindset: first understand the problem, then design a clean solution, and finally build something that can be maintained, reviewed and explained.
I am especially interested in developer-first products, process automation, fiscal integrations and internal tools that help teams work better.
Real client work
These are live client projects, not template mockups: strategy, development, technical SEO and conversion delivered end to end.
Local services · 5 languages
A 94-page service architecture designed to capture local searches and turn visits into quote requests.
Local business · mobile conversion
A clearer, faster route from a mobile visit to practical information and a booking decision.
Contact
It works best if you write with context, goal, rough timeline and the system or process you want to improve.