Ando Saabas
Principal ML Scientist Manager at Microsoft
Estonia
Summary
👤
Senior🎓
Top SchoolAndo Saabas is a Principal ML Scientist Manager with 11 years of experience blending machine learning, data science and software engineering to solve real-time media and large-scale product problems. He leads teams at Microsoft building deep-learning systems for call quality—noise suppression, echo cancellation and packet loss concealment—drawing on a long track record at Skype and Bolt where he helped stand up core ML services like ETA, dispatching and pricing. With a PhD in computer science and roots in research and infrastructure, he pairs rigorous statistical modeling with production-focused engineering. An active contributor to interpretability tooling, he authored core functionality in the treeinterpreter package to decompose tree-based model predictions into feature contributions. Based in Estonia, he is comfortable moving projects from research prototypes to robust, latency-sensitive deployments used by millions.
11 years of coding experience
17 years of employment as a software developer
PhD, Computer science, PhD, Computer science at Tallinn University of Technology
Stackoverflow
1,967reputation
194kreached
30answers
0questions
scikit-learn
top-5%
Github Skills (17)
python10
scikit10
machine-learning10
scikit-learn10
decision-tree10
random-forest10
numpy9
datasets9
data-analysis9
testing7
documentation7
bioinformatics6
classification6
pandas6
scipy6
Programming languages (2)
Github contributions (5)
andosa/treeinterpreter
Aug 2015 - Feb 2021
Role in this project:
Data Scientist
Data ScientistContributions:21 commits, 7 PRs, 19 pushes in 5 years 7 months
Contributions summary:Ando primarily contributed to the core functionality of the `treeinterpreter` package, focusing on interpreting scikit-learn decision trees and random forest predictions. They implemented and refined methods for calculating feature contributions, including conditional/joint contributions. The user also improved code quality through refactoring and documentation updates and added tests to validate the correctness of the prediction decomposition.
microsoft/AEC-Challenge
Sep 2020 - Jan 2023
AEC Challenge
Contributions:4 reviews, 15 commits, 8 pushes in 2 years 4 months
audio-classificationspectrogramaecspeech-enhancementvoice-activity-detection
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.