Projects

Here you can find some of the projects I have been involved in, along with a small description. Some of them also have a link to their own page.

Academic Projects

I have participated in many projects as part of my academic career. Here are the ones I am currently participating in:

• CompSLA (2024 - Present)
  The CompSLA group is an international cooperation group focused on developing tools and datasets for second language acquisition (SLA), with a special focus on lower-resourced languages. The group has organized two shared tasks, namely MultiGED-2023 on grammatical error detection and MultiGEC-2025 on grammatical error correction (as well as its associated dataset). The website for the working group can be found here.

• Mormor Karl (2023 - Present)
  Pseudonymisation can be a tricky affair. If done carelessly, we can end with phrases such as “grandma Karl is 27 years old”, which limits the usefulness of the data. The aim of this project is to make pseudonymisation in a way the both the original context is somewhat mantained but also that private information is protected. The website of the project can be found here.

Previous projects I have previously participated in are the following:

• Lund Rumour Mining Project (2022 - 2023)
  This project aims to study how rumours interact with vaccine hesitancy in Sweden. The websites of the project can be found here and here.

• COLMENA Misison (2015 - 2017)
  The objective of this mission was to send a small swarm of self-assembling robots to the moon. The payload was originally expected to launch mid-2021. The website for the project can be found here (in Spanish).

Reduce Soluciones

Reduce is a startup founded by me and my family in 2016. We are devoted to bringing renewable energy solutions to people in the central region of Mexico at an affordable price.

You can access the website (in Spanish) here or click the title of this section for more information.

Shared Tasks

I have been a co-organizer of the shared tasks below.

• MultiGEC-2025
  A shared task focused on multilingual grammatical error correction. It covers 12 European languages, namely Czech, English, Estonian, German, Greek, Icelandic, Italian, Latvian, Russian, Slovene, Swedish, and Ukrainian and consists of 17 subcorpora. It consisted on two tracks, one for minimal edits and another one that allowed fluency edits. The results were presented at the NLP4CALL workshop on March 5th, 2025. You can find the website for the shared task here and the website for the dataset here.
  Note: The MultiGEC dataset is meant to be updated with more languages and more subcorpora in the future. If you want to compare your results with those from the shared task, make sure that you are using the original release or one that is compatible with it.

Hackathons

Here you can find a list of the hackathons I have participated in.

• Code Against Hate (September 2020; Online)
  The idea was to propose a solution to help curb online hate speech without incurring in censorship. My team got first place by proposing hate.no, a grammarly-like plugin/API for messaging services. It would alert the user if the message they are about to post could be considered abusive. This allows the user to either change their message or to post it as-is. The idea is “think twice before you post”. You can find the presentation here.

• Hello World 2022 (February 2020; Edinburgh, Scotland)
  My team was the runner-up in the Quantum Black challenge of the hackathon. The idea was to use the kedro python package to propose an ML solution. We designed a model to predict whether an epidemic risks becoming a pandemic given flight information of the affected areas.