Projects

I’ve been involved in several research projects with both research groups and companies R&D. I am one of many coordinators of the applied research laboratory LAPPIS. I had the opportunity to coordinate several applied research projects with governmental agencies and private companies. I had more than 50 scholarships under my supervision. Most of the projects I am involved in have software development under an open source license. I will present the latest projects. For the complete list, please visit my lattes page.

A presentation of the latest projects, with some technical details is available here

Big Open Source Siblings (BOSS) - 2020 - ongoing

The BOSS initiative aims to reach underrepresented groups in tech who want to be mentored and assisted to learn how to work on open source projects. We will provide a safe space and an accompanying methodology for them to learn basic concepts of working in a small open source project in practice and build enough background for them to, at the end of the program, be able to choose other projects to continue their contributions. The mentoring process helps build confidence, fill in technical gaps, and engage the participants.

In the BOSS program, besides the technical aspects of introducing newcomers to open source, we deal with some challenges to inclusion, namely, intersectionality and impostor syndrome. The BOSS initiative manual is available here.

We won the Gnome Community Engagement Challenge.

image

Machine Learning for outlier detection in Customer databases - partnership with the Instituto Eldorado/Dell - 2021 - ongoing

In this project, we experiment with several machine learning algorithms to detect outliers and anomalies in large Customer Dataset. We deploy Data, Feature Engineering, and MLOps techniques and practices.

We adopted a commonly used machine learning workflow depicted in various forms across industry and research. It has commonalities with prior workflows defined in data science and data mining, such as TDSP, KDD, and CRISP-DM. They have in common the data-centered essence of the process and the multiple feedback loops among the different stages. We followed some DevOps and MLOps best practices and automation to speed up the experimentation cycle and scale the process to other databases and a larger team.

Chatbot to guide the submission of Culture Incentive proposal - Ministry of Culture - 2017-2019

In this project, we developed one of the first chatbots for brazilian government. We developed an open source chatbot solution, with over 70 intentions to answer about the Culture Incentive Law to guide cultural agents in proposing their project.

We innovate in chatbot architecture, continuous deploy pipeline, and we used MLOps practices and techniques before the term was coined by practitioners.

The entire project is available on the github organization of LAPPIS.

We also developed an Open Source Chatbot Boilerplate, with all the functionalities needed for a chatbot product, to guide newcomers into this growing industry. The community has more than 350 participants on the communication channel.

Public Money Public Code – Modernising Public Infrastructure with Free Software

The LAPPIS team translated to Brazilian Portuguese the publication “Public Money Public Code – Modernising Public Infrastructure with Free Software” from Free Software Foundation Europe. The translation in portuguese - Dinheiro Público Código Público - Modernizando a Infraestrutura Pública com Software Livre

Other Partnerships

The image illustrates some of the past/current LAPPIS partners.

parcerias