Duygu Altinok
Machine Learning Engineer at Palo Alto Networks
Berlin, Germany
Summary
👤
Senior🎓
Top SchoolDuygu Altinok is a Machine Learning Engineer based in Berlin with 11 years of experience specializing in ASR and speech-language systems, spanning HMM-GMM and MFCC-era acoustics through LSTMs to modern Transformer-based and foundational ASR models. She has driven state-of-the-art ASR research and production work—from on-device acoustic training and CTC decoding to entity-aware and semantics-infused speech models—while contributing to spaCy, Hugging Face and PyTorch ecosystems and authoring the bestseller "Mastering spaCy." Her work on Turkish NLP is both commercial and open-source, including substantive improvements to spaCy’s Turkish support and practical techniques for handling Turkish morphology. Known for blending high-level semantic ambitions with gritty engineering (she still loves a good grep), she also experiments with audioLMs and efficient attention implementations.
11 years of coding experience
6 years of employment as a software developer
Master’s Degree Mathematics, Master’s Degree Mathematics at Bilkent UniversityDoctor of Philosophy (Ph.D.) CS and Applied Mathematics, Doctor of Philosophy (Ph.D.) CS and Applied Mathematics at The University of BonnBachelor’s Degree Computer Engineering, Bachelor’s Degree Computer Engineering at Orta Doğu Teknik Üniversitesi / Middle East Technical University
English, Turkish, German, Hebrew
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (7)
tokenize10
tokenizer10
nlp10
spacy10
turkish10
python10
morphology9
Programming languages (6)
Github contributions (5)
explosion/spaCy
Nov 2017 - Jan 2022
💫 Industrial-strength Natural Language Processing (NLP) in Python
Role in this project:
Data Scientist
Data ScientistContributions:31 reviews, 39 commits, 37 PRs in 4 years 2 months
Contributions summary:Duygu primarily contributed to improving the Turkish language support within the spaCy library. Their commits involved cleaning and correcting encoding problems in Turkish stop words, adding example sentences for Turkish, ensuring alphabetical order for Turkish characters, and removing number words to carry into the lexical. Further contributions included adding commonly used cases, adding like\_num to the lexical, adding abbreviations, and adding Turkish morphological rules. These changes indicate a focus on enhancing the NLP capabilities of spaCy for the Turkish language.
fairness-mlpythondata-preprocessinglanguage-processingtokenization
DuyguA/spaCy
Feb 2018 - Dec 2024
💫 Industrial-strength Natural Language Processing (NLP) with Python and Cython
Contributions:163 pushes, 63 branches in 6 years 11 months
nlporangepythonlanguage-processingindustrial
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.