Alberto Bietti
Research Scientist, Flatiron Institute
New York, New York, United States
Summary
👤
Senior🎓
Top SchoolAlberto Bietti is a research scientist at the Flatiron Institute with 14 years of experience at the intersection of machine learning, optimization, and AI, combining academic rigor with industry impact from roles at Meta, NYU, Inria and Quora. His work spans foundational research and practical algorithm engineering—highlighted by contributions to Vowpal Wabbit that extend contextual bandit and cost-sensitive capabilities for production-scale online learning. Trained in applied mathematics and computer vision (PhD and graduate work across Grenoble, ENS, Mines de Paris and Caltech), he moves fluidly between theory and systems, prototyping methods that scale. A former visiting researcher at FAIR and faculty fellow at NYU, he brings deep expertise in online and interactive learning, often optimizing algorithms for real-world ranking and exploration problems. Notably, his profile reflects a blend of high-impact open-source tinkering and formal research, making him adept at turning advanced ML ideas into deployable systems.
14 years of coding experience
5 years of employment as a software developer
MPSI, MP*, Math, Physics, Engineering, Computer Science, MPSI, MP*, Math, Physics, Engineering, Computer Science at Lycée Louis Le Grand, ParisMaster of Science (MS), Machine Learning, Vision, Mathematics, Master of Science (MS), Machine Learning, Vision, Mathematics at Ecole normale supérieureDoctor of Philosophy - PhD, Applied Mathematics, Doctor of Philosophy - PhD, Applied Mathematics at Université Grenoble AlpesGraduate, Computer Vision, Graduate, Computer Vision at California Institute of TechnologyMaster of Science - MS, Engineering, Applied Mathematics, Master of Science - MS, Engineering, Applied Mathematics at Ecole nationale supérieure des Mines de Paris
English, French, Italian
Github Skills (12)
machine-learning10
bandit10
c-language10
cpp10
cprogramming-language10
elearning10
algorithms9
data-structure9
algorithm9
data-structures9
active-learning8
reinforcement-learning8
Programming languages (4)
Github contributions (5)
VowpalWabbit/vowpal_wabbit
Aug 2017 - Feb 2019
Vowpal Wabbit is a machine learning system which pushes the frontier of machine learning with techniques such as online, hashing, allreduce, reductions, learning2search, active, and interactive learning.
Role in this project:
ML Engineer
ML EngineerContributions:8 commits, 8 PRs, 19 comments in 1 year 6 months
Contributions summary:Alberto primarily focused on modifications and enhancements within the Vowpal Wabbit codebase related to Contextual Bandit (CB) algorithms. Their contributions involved implementing and refining features such as support for the adf reduction in the cbify module, as well as introducing the `--greedify` option for bagging. Furthermore, the user made changes to the cover and mtr explore features, and extended CB functionality to cost-sensitive examples. These modifications suggest a focus on improving and extending the capabilities of Vowpal Wabbit for advanced machine learning problems.
machine-learningvowpal-wabbitc-plus-plusonline-learningcontextual-bandits
abietti/online_hmm
May 2014 - Apr 2015
Online EM algorithms for hidden Markov and semi-Markov models + applications to audio segmentation and clustering
Contributions:37 commits, 3 pushes in 11 months
audioclusteringmarkovmarkov-modelsegmentation
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.