Carlo Lucibello
Machine Learning Engineer at Bocconi University
Turin, Piedmont, Italy
Summary
🤩
RockstarCarlo Lucibello is a Machine Learning Engineer based in Turin, Italy, with 12 years of experience at the intersection of statistical physics and practical ML engineering. He has a strong open-source footprint in the Julia ecosystem, contributing core features and performance fixes to projects like Flux.jl and Zygote.jl and adding models and optimizers to Knet.jl. Carlo’s work spans model implementation (LeNet5, convolutional VAEs), optimizer design (Nesterov momentum), and auto-differentiation internals, reflecting both research-grade rigor and production-minded reliability. He also extended graph-processing tooling by adding multiple file format supports to LightGraphs.jl, showing versatility beyond pure ML. Comfortable optimizing for GPUs and test-driven development, he blends low-level numerical care with high-level model design. An intriguing thread through his profile is the fusion of statistical physics intuition with hands-on engineering, informing robust, well-tested ML tooling.
12 years of coding experience
Github Skills (45)
algorithm10
graph-algorithms10
convolutional-neural-networks10
optimizers10
testing10
variational-autoencoder10
machine-learning10
data-structure10
gradient10
automatic-differentiation10
deeplearning-ai10
deep-learning10
optimisation10
mnist10
flux10
Programming languages (14)
Github contributions (5)
FluxML/model-zoo
Feb 2020 - Dec 2022
Please do not feed the models
Role in this project:
ML Engineer
ML EngineerContributions:43 reviews, 62 commits, 61 PRs in 2 years 10 months
Contributions summary:Carlo implemented and refined machine learning models within the Flux framework, focusing on the classification of MNIST data using convolutional neural networks (LeNet5). Their contributions included the creation of a LeNet5 model, defining data loading and preprocessing procedures, as well as writing training and evaluation functions. The user also integrated logging with TensorBoard and explored the use of CUDA for GPU acceleration, optimizing for performance.
deep-learningmachine-learningfluxfeedjulia
FluxML/Flux.jl
Oct 2017 - Jan 2023
Relax! Flux is the ML library that doesn't make you tensor
Role in this project:ML Engineer
ML EngineerContributions:573 reviews, 320 commits, 479 PRs in 5 years 3 months
Contributions summary:Carlo's contributions primarily focused on enhancing the Flux.jl library, specifically by adding new features and improving existing functionalities related to machine learning models. The user implemented dropout layers for regularization and introduced improvements to the optimizer implementations. Furthermore, the commits included the addition of a new layer and the implementation of various tests to ensure the reliability of the code.
ml-librarythe-human-braindata-sciencedeep-learningneural-networks
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.