Experience

Case studies

Three projects told by the problem they had to solve, not by my position on the org chart.

1. Multi-tenant IoT platform with real-time data

   Wireless DNA · 2024 — 2026

   Context

   An IoT product company with several clients on the same platform: sensors deployed in the field sending telemetry continuously, and customers who need to see it and act on it.

   Problem

   Data arrived from the device network (Chirpstack, MQTT), but every customer needed their own space, with their devices and their data separated from everyone else's — without multiplying either the codebase or the deployment per customer.

   Solution

   Full-stack applications with Node, NestJS and React, time series in InfluxDB and the rest of the model in MongoDB and MySQL. All containerised and deployed to the cloud with a multi-tenant design starting at the database, not patched on at the end.

   Result

   A single codebase and a single deployment serve every customer with isolated data, and the dashboards show telemetry in real time instead of after-the-fact reports.

   • Node.js
   • NestJS
   • React
   • InfluxDB
   • MongoDB
   • MQTT
   • Chirpstack
   • Docker

2. In-house infrastructure for a small company's dev team

   Solution Data Technology · 2023 — 2024

   Context

   A small technology company with its own development team, internal products and custom work for clients, spread across several stacks (MEVN, MERN and PHP).

   Problem

   There was no common way to deploy: every application was published its own way, and every integration with an ERP or a data source was solved from scratch again on the next project.

   Solution

   An in-house containerised cloud architecture for the company's services and applications, the development team coordinated around it, and the recurring ERPs and data sources integrated once.

   Result

   Deployment stopped depending on who was doing it: both new and inherited applications moved onto the same infrastructure, and the team worked against an environment that matched production.

   • Docker
   • Node.js
   • Vue
   • React
   • PHP
   • Cloud

3. Integrating heterogeneous travel suppliers

   Conecta Turismo · 2022 — 2023

   Context

   A travel-sector product fed by the catalogues of several external suppliers, each with its own API, its own format and its own update cadence.

   Problem

   Some suppliers exposed REST and others SOAP, with data models that looked nothing like each other. Every new integration threatened to become one more special case inside the product.

   Solution

   Integrations against the suppliers' APIs and ETL pipelines collecting the different travel products with Node.js, Go and MongoDB on Google Cloud Platform, normalising whatever comes in into a single internal model.

   Result

   The product queries one data model, and the difference between a SOAP supplier and a REST one never reaches the application layer.

   • Node.js
   • Go
   • MongoDB
   • Google Cloud Platform
   • REST
   • SOAP