Silvia Terragni
Senior Lead ML Engineer at Upwork
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSilvia Terragni is a Senior Lead ML Engineer based in San Francisco with 11 years of experience bridging academic research and production ML, currently leading ML efforts at Upwork after roles in research-heavy startups and academia. She holds a Ph.D. in Computer Science focused on topic modeling and has tangible impact on influential open-source projects—contributing to contextualized-topic-models, OCTIS, and core Gensim improvements that sharpen evaluation and coherence metrics. Silvia combines deep probabilistic modeling expertise with hands-on engineering, improving ETM/LDA implementations and evaluation pipelines to make topic models more reliable in real-world systems. A pragmatic researcher, she has a track record of shipping reproducible tools from EACL/ACL-published work into widely used Python packages, and she often focuses on the subtle but crucial details of evaluation and model behavior.
11 years of coding experience
7 years of employment as a software developer
Master's degree, Computer Science, 110/110 cum Laude, Master's degree, Computer Science, 110/110 cum Laude at Università degli Studi di Milano-BicoccaLiceo Statale Scientifico e Classico "Ettore Majorana"
Italian, English, Spanish
Stackoverflow
176reputation
6kreached
8answers
0questions
Github Skills (21)
algorithms10
pytorch10
gensim10
python10
evaluation10
machine-learning10
topic-models10
natural-language-processing10
topic-modeling10
nlp10
metric10
testing9
bert9
scikit8
scikit-learn8
Programming languages (2)
Github contributions (5)
MIND-Lab/OCTIS
Mar 2020 - Jan 2023
OCTIS: Comparing Topic Models is Simple! A python package to optimize and evaluate topic models (accepted at EACL2021 demo track)
Role in this project:
Back-end Developer
Back-end DeveloperContributions:10 reviews, 640 commits, 16 PRs in 2 years 10 months
Contributions summary:Silvia appears to be primarily focused on refactoring the topic models, and improving the implementation of different functions such as retrieving word topic weights. They modify and update the core components of the topic models. The user's work is focused on improving the performance of the ETM and LDA models by adding code improvements.
pythonneural-topic-modelstopic-modelingbayesian-optimizationnatural-language-processing
MilaNLProc/contextualized-topic-models
Apr 2020 - Oct 2022
A python package to run contextualized topic modeling. CTMs combine contextualized embeddings (e.g., BERT) with topic models to get coherent topics. Published at EACL and ACL 2021 (Bianchi et al.).
Role in this project:Data Scientist
Data ScientistContributions:3 reviews, 82 commits, 1 PR in 2 years 6 months
Contributions summary:Silvia's commits primarily involve modifications to the `contextualized-topic-models` package, specifically focusing on evaluation metrics and model behavior. They introduced and fixed various aspects of the evaluation measures, including coherence and topic diversity, as well as the handling of topic distributions. The contributions focused on improving the model's functionality and performance analysis.
pythontext-as-databertneural-topic-modelstopic-modeling
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.