Marco Cognetta
Senior Software Engineer at Google
Tokyo, United States
Summary
👤
Senior🎓
Top SchoolMarco Cognetta is a Senior Software Engineer and PhD student specializing in natural language processing, with 11 years of experience building language models, federated learning pipelines, and interpretable ML systems at Google and in academia. He combines production-grade engineering for Gboard with research on tokenization, federated analytics, and model compression, and has published work on finite-state language model compression. Comfortable across languages and stacks, he contributes to high-profile open-source projects (Julia core, Flux.jl, NetworkX) implementing robust data structures, graph generators, and RNN examples. His background in formal languages and automata theory informs practical optimizations for language modeling and efficient parsing algorithms. Based in Tokyo with prior teaching and curriculum design experience, he brings a rare mix of deep theoretical training and hands-on product impact.
11 years of coding experience
7 years of employment as a software developer
Non-Degree-Seeking Graduate Student Scientific Computing, Non-Degree-Seeking Graduate Student Scientific Computing at Florida State UniversityDoctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Tokyo Institute of TechnologyBS Discrete Mathematics; Korean (Minor), BS Discrete Mathematics; Korean (Minor) at Georgia Institute of TechnologyMS Theoretical Computer Science, MS Theoretical Computer Science at Yonsei University
English, Korean
Github Skills (30)
unit-testing10
graph-algorithms10
matrix10
python10
artificial-neural-networks10
testing10
machine-learning10
rnn-model10
n10
data-structure10
mat10
deeplearning-ai10
deep-learning10
heap10
numerical-methods10
Programming languages (12)
Github contributions (5)
FluxML/model-zoo
Sep 2019 - Dec 2022
Please do not feed the models
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 8 commits, 5 PRs in 3 years 3 months
Contributions summary:Marco focused on modifying and improving the `char-rnn.jl` example within the repository, which appears to be a character-level recurrent neural network. The user's contributions include refactoring code, addressing deprecated methods, and adding new features like seed functionality and epochs to the training process. They also made updates to align with newer versions of dependencies.
deep-learningmachine-learningfluxfeedjulia
JuliaCollections/DataStructures.jl
Sep 2018 - Oct 2019
Julia implementation of Data structures
Role in this project:Backend Developer / Data Structures Engineer
Backend Developer / Data Structures EngineerContributions:5 commits, 6 PRs, 5 comments in 1 year
Contributions summary:Marco primarily focuses on implementing and refining data structures in Julia, specifically working on heaps, disjoint sets, tries, queues, stacks, deques, circular buffers, and linked lists. Their contributions include fixing bugs, adding new features like type conversion and the `eltype` function, and incorporating equality checks. The commits demonstrate a deep understanding of data structure implementation and testing in the context of a Julia project.
data-structuresimplementation-of-data-structuresdatastructuresjulia
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.