Francesco Visin
Staff Research Scientist at Google DeepMind
United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolFrancesco Visin is a Senior Research Scientist at Google DeepMind with 11 years of experience building and optimizing deep learning models, currently focused on Gemma models to democratize access to frontier language models. He combines a strong academic foundation—a PhD from Politecnico di Milano and research stints at MILA and LISA—with hands-on contributions to core ML tooling, including performance-sensitive work on Theano/PyTensor and maintenance of influential libraries like pylearn2 and Blocks. His background spans research, engineering, and teaching, having tutored large undergraduate cohorts and designed practical software for industry early in his career. Francesco’s contributions often blend systems-level optimization with careful engineering hygiene—adding C-level kernels, tests, and clearer documentation—to make research code robust and reusable. Colleagues value him for shipping production-ready research and for quietly improving developer experience in foundational open-source projects.
12 years of coding experience
13 years of employment as a software developer
Bachelor of Science (B.Sc.), Engineering/Industrial Management, Bachelor of Science (B.Sc.), Engineering/Industrial Management at Politecnico di MilanoMaster of Science (M.Sc.), Computer Science, Master of Science (M.Sc.), Computer Science at Università degli Studi di Milano-Bicocca
English, German, Italian
Github Skills (13)
neural-network10
machine-learning10
convolutional-neural-networks10
deep-learning10
python10
cudnn10
theano10
documentation10
c119
c179
api-documentation9
pep8
linear-algebra7
Programming languages (10)
Github contributions (5)
Theano/Theano
Sep 2015 - Nov 2016
Theano was a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. It is being continued as PyTensor: www.github.com/pymc-devs/pytensor
Role in this project:
Back-end Developer
Back-end DeveloperContributions:43 commits, 28 PRs, 12 pushes in 1 year 3 months
Contributions summary:Francesco contributed to the Theano library by implementing and testing the `LogSoftmax` activation function, along with optimizations. Their work included adding the Python code, tests, and C code. The user also addressed issues related to data type handling and test value printing. Additionally, the user made several optimizations to enhance the performance of the library.
mathematical-expressionsmultidimensional-arrayspymcpythontheano
mila-iqia/blocks
Mar 2015 - Feb 2016
A Theano framework for building and training neural networks
Role in this project:Technical Writer
Technical WriterContributions:6 commits, 6 PRs, 86 comments in 11 months
Contributions summary:Francesco primarily contributed to improving the documentation of the Blocks framework. Their work involved fixing docstrings, adding missing documentation for various modules, and correcting typos throughout the documentation. They also updated tutorial content and API references to enhance the clarity and usability of the framework's documentation. Furthermore, the user added a test case for error handling in `VariableFilter`.
neural-networktheano
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.