Ryan Serrao
Senior Machine Learning Engineer at Adobe
Cambridge, Massachusetts, United States
Summary
👤
Senior🎓
Top SchoolRyan Serrao is a Senior Machine Learning Engineer with a decade of experience building production ML systems at Adobe and Microsoft and a current MS Data Science candidate at Columbia University. He blends research-driven NLP and explainability work—contributing to the widely used SHAP library by integrating transformer-based text generation explainers—with hands-on engineering for products like Cosmos DB pricing, Whiteboard ink detection, and personalized recommendations. His background spans academic NLP research in Germany and India, practical large-scale data pipelines (extracting skills from 70M resumes), and exploratory soft-computing work on fuzzy cognitive maps for forest-fire prediction. Based in Cambridge, MA, he pairs strong systems and backend skills (Spark, PyTorch, Databricks) with a track record of open-source impact that bridges model interpretability and transformer integration.
10 years of coding experience
3 years of employment as a software developer
Bachelor of Technology - BTech Computer Science, Bachelor of Technology - BTech Computer Science at Vellore Institute of TechnologyMaster of Science - MS Data Science, Master of Science - MS Data Science at Columbia University
French, Hindi, English
Github Skills (8)
transformers10
text-generation10
pytorch10
machine-learning10
explainable-artificial-intelligence10
nlp10
python10
batch-processing8
Programming languages (1)
Github contributions (5)
shap/shap
Nov 2020 - Feb 2021
A game theoretic approach to explain the output of any machine learning model.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1 review, 220 commits, 16 PRs in 2 months
Contributions summary:Ryan's commits primarily focused on implementing and refining machine learning models for text generation within the SHAP library. They added a `Model` base class and then defined a `PTTeacherForcingLogits` class that integrates with PyTorch models for computing log odds. Furthermore, the user integrated a text generation model into the framework, encompassing various functionalities like text infilling and batch processing, indicating a focus on model explainability and integration of transformer models.
explaininterpretabilityshapdeep-learningapproach
ryserrao/shap
Sep 2020 - Feb 2021
A game theoretic approach to explain the output of any machine learning model.
Contributions:122 pushes, 17 branches in 5 months
approachexplainmachine-learning
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.