Alexandre Passos
Member Of Technical Staff at Periodic Labs
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolAlexandre Passos is a Member of Technical Staff based in San Francisco with 19 years of experience building large-scale machine learning systems and production software. He has held senior engineering roles at Google and OpenAI and now contributes at Periodic Labs, blending research rigor from his MSc/PhD training with pragmatic product delivery. Alexandre is an active open-source contributor to foundational ML projects—most notably improving scikit-learn’s SVD and mixture model implementations—and has deep experience in probabilistic modeling and inference from work on FACTORIE. He’s skilled at optimizing algorithms for large, sparse matrices and debugging complex EM and mean-field inference pipelines, bringing both implementation finesse and clear documentation to projects. Colleagues rely on him for turning advanced ML research into reliable, performant production components.
19 years of coding experience
11 years of employment as a software developer
BS Computer Science, BS Computer Science at Universidade Federal da BahiaUniversity of Massachusetts Amherst
Portuguese, English
Github Skills (20)
algorithm10
algorithms10
python10
scikit10
machine-learning10
pca10
parameter-tuning10
inference10
optimisation10
scikit-learn10
svd10
linear-algebra10
optimization10
data-science9
machine-learning-algorithms9
Programming languages (11)
Github contributions (5)
factorie/factorie
Oct 2012 - Jan 2014
FACTORIE is a toolkit for deployable probabilistic modeling, implemented as a software library in Scala. It provides its users with a succinct language for creating relational factor graphs, estimating parameters and performing inference.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:884 commits in 1 year 2 months
Contributions summary:Alexandre made significant contributions to the factorie/factorie project, with a focus on enhancing the inference capabilities and model functionality. Their work included debugging and improving the mean-field inference, caching variable sets, and fixing issues within the Expectation Maximization (EM) algorithm. They also contributed to model training and parameter optimization, implementing features related to linear algebra and Tensor2s. The user demonstrated the ability to identify and resolve bugs.
library-softwareparametersrelationaldeployableprobabilistic-modeling
scikit-learn/scikit-learn
Nov 2010 - Oct 2012
scikit-learn: machine learning in Python
Role in this project:Data Scientist & ML Engineer
Data Scientist & ML EngineerContributions:87 commits in 1 year 11 months
Contributions summary:Alexandre contributed significantly to the scikit-learn library by adding random projections SVD to scikits.learn.pca, enabling faster computation for large matrices. Further work included adding a power iteration parameter to the fast_svd implementation, likely to improve performance on high-rank, sparse matrices. The user's commits also included adding and improving the documentation for the Dirichlet Process Gaussian Mixture Model (DPGMM) and Variational Bayesian Gaussian Mixture Model (VBGMM) classes, suggesting a focus on model improvements and documentation.
data-analysispythonstatisticsdata-sciencelearn-machine-learning
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.