Skip to content
Raúl Cerón
Digital development & consulting

Backend engineering, delivered end to end.

I design and build the systems businesses actually run on — ERP customisation, commerce platforms, integrations and data infrastructure — and carry them through to the web and mobile apps people use every day.

Based in Zaragoza. Working with clients in Spain, Mexico and the United States.

Services

What I build

Six areas that come up again and again. Most projects combine three or four of them.

01

ERP & Odoo development

Custom Odoo modules, migrations off legacy in-house ERPs, and the sales, purchasing, helpdesk and reporting flows a business runs on. Delivered through a feature → dev → main release flow with client UAT before anything reaches production.

Odoo 18 Python XML-RPC PostgreSQL
02

Headless commerce

Strapi-backed APIs with Nuxt storefronts, and customisation of existing WooCommerce estates. Catalogue import tooling, variant and attribute logic, shipping rules by postal code, and payment gateway integration.

Strapi Nuxt WooCommerce Payment gateways
03

APIs & integrations

REST and GraphQL services, webhook pipelines, and the unglamorous connective work between systems that were never designed to talk to each other — payment providers, logistics, public data sources, legacy ERPs over XML-RPC.

REST GraphQL Webhooks XML-RPC
04

Databases & search

Schema design, multi-tenant isolation enforced with row-level security at the database rather than in application code, migrations with reconciliation, and semantic search built on pgvector embeddings alongside your relational data.

PostgreSQL pgvector RLS NoSQL
05

Cloud architecture

Infrastructure as code across OCI, AWS and GCP: containerised deployments, per-tenant storage and key management, network and policy gates, and the boring reliability work that decides whether a system survives its second year.

OpenTofu OCI AWS GCP Docker
06

AI integration

Embedding pipelines, retrieval and classification wired into systems you already have — not a chatbot bolted onto the side. Including the harder question of what data is allowed near a model, and how to prove it never got there.

Embeddings RAG LLM pipelines
Selected work

A few recent projects

Client names are withheld. The technical detail is accurate.

ERP migration for an industrial manufacturer

Mexico

Replaced a legacy in-house ERP with Odoo 18, building roughly nineteen custom modules across helpdesk, sales, purchasing and market intelligence. Released through a gated branch flow with client acceptance testing at each block. Data migration ran under a strict no-exposure protocol: real records stayed quarantined outside the development environment and only synthetic fixtures were used in code.

Odoo 18 Python PostgreSQL Docker nginx

Zero-knowledge board governance platform

In development

A document platform for board directors where the server never needs plaintext. End-to-end encrypted document vault, append-only hash-chained audit log, cryptographic resolution voting, and an opt-in searchable tier using vector embeddings. Tenant isolation is enforced by the database with row-level security, not by application logic.

FastAPI SQLAlchemy 2 PostgreSQL 18 + pgvector React Expo OpenTofu

Headless commerce platform

Mexico

Strapi API and Nuxt storefront for a distributor, with catalogue import tooling, shipping-zone resolution by postal code across multiple carriers, and payment gateway integration. Included the migration and clean-up of an inherited product and pricing dataset.

Strapi Nuxt PostgreSQL GCP

Market intelligence collector

Mexico

A standalone collector that pulls public trade and regulatory sources — news feeds, the US Federal Register, Census trade data, Banxico and FRED indicators — classifies and summarises them in Spanish with an LLM, and pushes the results into the client ERP over XML-RPC. Python standard library only, no dependencies.

Python LLM classification XML-RPC Public data APIs
How I work

Three things that do not change

Backend first

The data model and the security boundaries get settled before anything is drawn. Most of the expensive problems in a project are decided in the first week, in the schema.

Tests before implementation

No feature without a failing test first. Small commits, gated releases, and acceptance testing with you before anything reaches production.

AI-assisted, not AI-generated

My background is backend. Assisted tooling is what lets me carry a project across the whole stack — architecture through to web and mobile front ends — but everything that ships has been read and understood.

About

Who you would be working with

I am a backend engineer working independently from Zaragoza, with clients across Spain, Mexico and the United States. Most of my work is with established businesses that have a system doing something important and need it extended, replaced or connected to something else.

That means I spend more time on data models, integration boundaries and deployment than on greenfield product design. When a project does need a front end — an admin panel, a storefront, a mobile app — assisted tooling lets me take it there myself rather than handing off half a system.

Contact

Tell me what you are building

A short description of the system and the problem is enough to start. I reply to everything within two working days.

hola@raulceron.es
Available for new projects Invoicing in EUR or USD · Spain, Mexico and the United States