Projects
Things I've built, alone or with good people.
- 2026
AnxietyCode ↗
Anxiety is a tool I built to stop piling up digital junk on my Mac.
More +Less −
It's not a polished product, it's just my personal fix for the mental friction that comes from having a chaotic Downloads folder. It comes from a simple mental framework: identify what's junk (.dmg installers, .app files, single-use files), reduce the emotional friction of deleting them by moving them to a transition zone instead of deleting them outright, and automate the habit, knowing that my psychological resistance to deleting things is stronger than my discipline. Right now it uses a file watcher that monitors the Downloads folder in real time. When a new file arrives, it applies automatic rules: it detects installers, moves them to should-be-deleted, avoids name collisions and keeps everything traceable without interrupting my flow. It doesn't delete anything directly, it just sets the stage so I can let go of it later without anxiety. It's built with Python, Typer, Watchdog and uv. It runs in the foreground or as a background service using a Launch Agent on macOS. I designed it for my own setup, but if it inspires your own automations, even better. Going forward I want to package it on PyPI, create a Homebrew formula, extend it to the Mac desktop and add auto-purge rules to delete old files without me having to think about it. It's small, but it's about reclaiming mental clarity one file at a time.
python
- 2025
NotMeta
NotMeta was born with the idea of transforming how companies manage WhatsApp Business, reusing my previous experience building WhatsApp-based products and several tools I'd created…
More +Less −
NotMeta was born with the idea of transforming how companies manage WhatsApp Business, reusing my previous experience building WhatsApp-based products and several tools I'd created in personal projects, like a scheduling app or staff management systems. Starting a project like this in my free time forced me to make the most of components I'd already built and to choose technologies that let me move fast without giving up quality. The motivation came from noticing that, in many organizations, customer conversations are concentrated on a single phone, which makes it hard to share the work, keep a centralized record and measure the team's efficiency. NotMeta was conceived to solve that problem: a platform that lets several employees connect to the same WhatsApp Business number and handle conversations simultaneously from an interface similar to the app they already know. During development we focused on creating a collaborative experience. We implemented a roles and permissions system, a metrics dashboard to analyze response times and performance, and integrations that let you download and review every conversation. All of this was done in compliance with the WhatsApp Business API policies, while aiming for the simplest possible setup for companies. The project is in advanced stages and will soon be piloted by two companies that will help us refine the functionality before its official launch. Taking part in startup programs from AWS, Google Cloud, Azure and GitHub has given us additional resources and support, and has let us validate the business model beyond the technology side. NotMeta's technical architecture reflects that mix of ambition and pragmatism. We went with Next.js for the web application, while Astro is used for the landing page and static materials, and a second Next.js project serves as a catalog of products and services. The heart of the business logic and the asynchronous tasks lives in Django and Celery. All the infrastructure is deployed on AWS, with services like Amplify and App Runner for the front‑end, a Fargate cluster for the backend services, Route 53 for DNS, ECR for container images and RabbitMQ for queue management, plus RDS with PostgreSQL for the database and SNS for email notifications. The continuous integration and delivery pipeline is handled through GitHub Actions, which builds the images and automates deployments. We've also built mobile apps with React Native and Expo to offer a universal experience on Android and iOS, and we use Tailwind, shadcn and Gluestack to build consistent, customizable interfaces. For internal automations we rely on n8n, and we've integrated JupyterLab with Django to run scripts and test concepts quickly. The artificial intelligence features are built on the LangChain, LangGraph and LangSmith libraries, which let us explore new forms of interaction and analysis. NotMeta has become one of my best projects to date and, without a doubt, the most ambitious. I've put into it all my technical knowledge, my leadership skills and what I've learned about business, always backed by the optimism and perseverance that entrepreneurship demands. Even if in the end it doesn't reach the commercial success we imagine, its modular design and its focus on WhatsApp features make it reusable in other personal projects; in fact, it was also born with the idea of serving as the foundation for several side projects. Throughout its development, NotMeta hasn't just been a technical challenge; we've also had to get through numerous bureaucratic processes and business validations specific to the companies we work with. Managing permissions, complying with policies and adapting the solution to legal requirements has taught us that a real project goes far beyond technology. All in all, working on NotMeta has been a continuous and rewarding learning experience. Combining multiple technologies and services, adapting them to a business messaging context, and at the same time aligning the product with the real needs of customer service has let us build a robust tool that's ready to scale. Reusing work from side projects and choosing a stack designed to move fast has been key to getting results in a limited amount of time. Now we can't wait to see how the first pilot companies use the platform and help take it to the next level.
- 2025
Damascuss real time feedbackLive ↗
Damascuss was a startup-style initiative that aimed to connect brands present in supermarkets and hypermarkets with instant visibility at the point of sale.
More +Less −
To do this, it used a network of on-demand collaborators who carried out scanning missions on shelves and other spaces, at the indirect request of the brands. Brands had a dashboard where they could configure missions with features typical of marketing campaigns (budget, frequency, duration, segments or zones, among others), which let them monitor in real time how their perfect store strategies, trade agreements and promotional campaigns were being executed. The project was ultimately cancelled and pivoted into a more ambitious initiative, since we weren't able to secure the funding essential to operate it. However, the infrastructure, the experience gained and the resources built were strategically reused in a new project.
AstroDjangoPostgresqlGithubGithub actionsAwsExpo react nativeNextjs
- 2018
CreatripLive ↗
Creatrip was a Software Factory–style venture dedicated to building custom solutions for different kinds of clients.
More +Less −
We built mobile and desktop applications, backend and frontend systems, database management and infrastructure, with a strong component of technology advisory and consulting. Even though the business was going well, for personal reasons I decided to go back to working at companies, since I had set myself the goal of getting into a big tech company like Google or Amazon at least once. I believed great projects were built as a team and felt I still had a lot to learn. I would reach that goal later on, after a tough but very rewarding path that paid off significantly in soft skills, technical knowledge, financial growth and valuable connections. On top of that, Creatrip gave me a business and technology perspective that went beyond the purely technical, which allowed me to set new goals from a much more mature professional profile.
PythonPhplaravelDjangoLumenMysqlSeleniumLinuxVuejsAngularJavaIonicTypescript