← andresvidal.dev

Andrés Vidal

Computer Scientist & Statistician

andres@andresvidal.dev andresvidal.dev github.com/andres-vidal linkedin.com/in/andresvidalberriel orcid.org/0009-0008-6350-7585 Montevideo, Uruguay

Uruguayan computer scientist and statistician working to bridge the gap between statistical machine-learning research and fast, reliable software. Peer-reviewed author in distributed generative models, and author of ppforest2, a high-performance C++/R implementation of Projection Pursuit Random Forests. 6+ years building production software for United States clients.

Education

Upcoming
MSc, Universidad de la República, Uruguay
Faculty of Engineering (machine learning / applied mathematics).
Ongoing
BSc in Statistics (Technology profile), Universidad de la República, Uruguay
Statistical learning, multivariate analysis, and mathematical modelling.
2022
Computer Science & Engineering (Ingeniería en Computación), Universidad de la República, Uruguay
Thesis: Aumentación de conjuntos de datos utilizando redes neuronales generativas profundas distribuidas. Advisor: S. Nesmachnow.
2015
Technical Diploma, IFRS Campus Canoas (Brazil)
Originally Técnico em Informática, with emphasis on systems analysis & development; the program was later renamed Técnico em Análise e Desenvolvimento de Sistemas.

Publications

2023
Parallel-Distributed Implementation of the Lipizzaner Framework for Multiobjective Coevolutionary Training of Generative Adversarial Networks
S. Nesmachnow, J. Toutouh, A. Mautone, G. Ripa, A. Vidal
High Performance Computing · CARLA 2023, Springer · doi.org/10.1007/978-3-031-52186-7_7
2023
Multiobjective coevolutionary training of Generative Adversarial Networks
A. Mautone, G. Ripa, A. Vidal, S. Nesmachnow, J. Toutouh
GECCO '23 Companion, ACM
2022
Aumentación de conjuntos de datos utilizando redes neuronales generativas profundas distribuidas: exploración del uso de algoritmos coevolutivos multiobjetivo en busca de mejoras en la diversidad de las muestras generadas
Computer Engineering thesis · with A. Mautone and G. Ripa · Advisor: S. Nesmachnow
Universidad de la República · Colibrí repository

Experience

Since 2020
Senior Software Engineer, WyeWorks
  • Full-stack development of production web apps for US clients in TypeScript (React, Next.js, NestJS), across ed-tech, e-commerce/ticketing, and healthcare.
  • Prisma (virtual school): real-time collaborative learning tools, including a collaborative slideshow editor with embedded video and a Notion-like block editor, built with React and CRDTs.
  • Led design of a schema-driven form-rendering framework (React + Zod) to replace a legacy enterprise forms system.
  • Designed and ran WyeWorks' internal onboarding program.
Since 2013
Research Assistant, IFRS Campus Canoas
Held Scientific Initiation research scholarships (CNPq, IFRS) on interdisciplinary projects that applied software to the humanities and education: Brazilian literature in translation (which grew into the Richard Burton platform) and educational robotics. Also taught occasional short courses (web technologies 2020; Elixir/Phoenix 2023).

Selected Projects

ppforest2
Projection Pursuit Random Forests: a high-performance implementation
A fast, memory-efficient C++ core with R and CLI interfaces, and a modern successor to the R PPforest package. Oblique splits, multi-threaded (OpenMP) training, tidymodels integration, and reproducible cross-platform golden tests. Submitted to CRAN. (Statistics thesis project.)
github.com/andres-vidal/ppforest2 · docs
richardburton
Richard & Isabel Burton Platform
A full-stack platform (React + Elixir/Phoenix) cataloguing English translations of Brazilian literature. Grown from a Java (JSF/JPA) technical-degree project into an INPI-registered platform, now an ongoing volunteer research collaboration with IFRS Canoas.
github.com/andres-vidal/richardburton

Skills

Programming
C++, R, TypeScript, Elixir, Python
Frameworks / tools
React, Next.js, NestJS, Zod, PostgreSQL · tidymodels, Phoenix · Git, CMake, OpenMP, CI/CD
Areas
Statistical & machine learning · high-performance / parallel computing · research software engineering · full-stack web
Languages
Spanish (native) · Portuguese (bilingual) · English (Cambridge FCE, C1)