Artidoro Pagnoni
Research Scientist at Meta
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolArtidoro Pagnoni is a PhD student in NLP at the University of Washington and a visiting researcher at Meta with 11 years of software engineering and research experience bridging large-scale ML systems and language-model fine-tuning. He earned an MS at CMU and BA/MS degrees from Harvard in Physics and Computer Science, and previously built ML tooling at Microsoft where he brought machine learning to C# via ML.NET. His open-source contributions span ML.NET refactors and TorchSharp I/O work to practical LLM finetuning tools like QLoRA, showing comfort from low-level data handling to cutting-edge LLM pipelines. He has a track record of converting legacy trainers to modern estimator patterns and enhancing finetuning scripts for benchmarks like MMLU, reflecting an emphasis on reproducibility and engineering rigor. Comfortable in both research and production contexts, he combines academic depth with pragmatic software development across industry labs and open-source projects. Based in Seattle, he blends physics-rooted quantitative thinking with hands-on ML systems engineering.
11 years of coding experience
8 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at University of WashingtonBaccalauréat Scientifique, Mathematics, Baccalauréat Scientifique, Mathematics at Lycée Chateaubriand de RomeMaster of Science - MS, Computer Science, Master of Science - MS, Computer Science at Harvard UniversityResearch Master, NLP, Research Master, NLP at Carnegie Mellon University
English, Italian, French, German
Stackoverflow
1reputation
0reached
2answers
0questions
Github Skills (26)
transformers10
pytorch10
python10
evaluation10
mldotnet10
machine-learning10
mlnet10
trainings10
fileio10
fine-tuning10
dotnet-core10
file-processing10
file-handling10
file-access10
csharp10
Programming languages (8)
Github contributions (5)
artidoro/qlora
May 2023 - Jul 2023
QLoRA: Efficient Finetuning of Quantized LLMs
Role in this project:
ML Engineer
ML EngineerContributions:17 reviews, 19 PRs, 45 pushes in 2 months
Contributions summary:Artidoro's commits primarily focus on modifying and improving the `finetune_llama.py` script, which suggests involvement in fine-tuning language models. Code changes include the addition of features related to model prediction, dataset handling, and evaluation metrics, specifically for the MMLU benchmark. The user also contributed to the project's overall structure by renaming and cleaning up core files, modifying configuration scripts, and adding examples.
fine-tuning
dotnet/machinelearning
Aug 2018 - Aug 2019
ML.NET is an open source and cross-platform machine learning framework for .NET.
Role in this project:ML Engineer
ML EngineerContributions:81 commits, 192 PRs, 73 pushes in 11 months
Contributions summary:Artidoro's commits primarily focused on substituting a wine quality dataset with a machine-generated dataset within the test suite. They modified several test files related to core ML components, predictors, and the testing framework to accommodate the new dataset. Further contributions involved converting existing trainers, specifically Random, Prior, OlsLinearRegression, Logistic, Multiclass Logistic, and Poisson Regression, to the standard estimator format within the ML.NET framework, indicating a focus on refactoring and standardizing training components.
dotnetmachine-learningml
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.