Ekin Akyürek
Member Of Technical Staff at OpenAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolEkin Akyürek is a machine learning researcher and engineer with a decade of experience bridging cutting-edge research and production, currently a Member of Technical Staff at OpenAI after completing a PhD at MIT CSAIL. His work focuses on compositional generalization in neural language models and using language to guide few-shot learning for image classifiers and robot policies, with internships at Google studying in-context learning and data attribution. An early contributor to Knet.jl, he has practical deep learning engineering chops—implementing model serialization, fixing low-level bugs, and extending loss/softmax functionality—complementing his academic rigor. He has co-founded startups that shipped ML-driven customer engagement products and has a background spanning embedded systems, full-stack development, and numerical computation, reflecting a rare blend of systems, research, and entrepreneurship. Based in San Francisco, he combines principled research with hands-on implementation, often surfacing simple inductive biases to unlock stronger generalization in large models.
10 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Massachusetts Institute of TechnologyBachelor’s Degree Electrical and Electronics Engineering, Bachelor’s Degree Electrical and Electronics Engineering at Koç University
English, Turkish
Stackoverflow
59reputation
2kreached
2answers
3questions
Github Skills (19)
machine-learning10
data-serialization10
serialization10
deep-learning10
neural-network10
knexjs10
knn10
julia10
n9
rnn-model9
data-science8
cordova-plugin6
swift6
uitableview6
ionic-framework6
Programming languages (7)
Github contributions (5)
denizyuret/Knet.jl
Aug 2017 - Jun 2020
Koç University deep learning framework.
Role in this project:
ML Engineer
ML EngineerContributions:23 commits, 21 PRs, 25 pushes in 2 years 10 months
Contributions summary:Ekin contributed to the Knet.jl deep learning framework by implementing a JLD2 interface for saving and loading models, including RNNs and KnetArrays, enabling deep serialization. They fixed a unicode character error in a charlm example and debugged the resnetlib, correcting an issue with loading moment variances. They also added functionality related to softmax, logsoftmax, logistic loss, and binary-cross-entropy, improving the framework's capabilities.
data-sciencedeep-learningknetneural-networksmachine-learning
ekinakyurek/Morse.jl
Nov 2018 - Nov 2019
Paper Title: Morphological analysis using a sequence decoder. Package Summary: Neural Morphological Analyzer + Lemmatizer
Contributions:43 commits, 2 PRs, 20 pushes in 11 months
nlpsequencemorphologicaldecodernatural-language-processing
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.