Kejsi Take
Senior Security Engineer at Meta
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolKejsi Take is a Senior Security Engineer at Meta with nine years of experience at the intersection of security, privacy, and machine learning. She recently defended a PhD at NYU Tandon on improving security and privacy by analyzing affordances for online harassment, blending rigorous research with practical mitigation strategies. Her background includes ML and data science internships at Twitter and Microsoft and hands-on open-source contributions to sktime, where she improved time-series forecasting components and quantile prediction. At NYU she taught courses on privacy and machine learning, reflecting strong communication skills and an ability to translate research into curriculum. Kejsi brings both academic depth and production engineering experience, focusing on measurable improvements to user safety and model robustness. Colleagues describe her as a researcher-engineer who surfaces subtle affordances attackers exploit and turns those insights into deployable defenses.
9 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at New York UniversityB.A. in Computer Science, Information Systems, B.A. in Computer Science, Information Systems at American University in BulgariaStudy Abroad, Computer Science, Study Abroad, Computer Science at Truman State University
English, French, Italian, Albanian
Github Skills (11)
scikit10
forecasting10
pandas10
machine-learning10
time-series10
forecast10
python10
scikit-learn10
testing9
prophet9
anomaly-detection5
Programming languages (5)
Github contributions (5)
sktime/sktime
Oct 2021 - Mar 2022
A unified framework for machine learning with time series
Role in this project:
Data Scientist
Data ScientistContributions:7 reviews, 7 commits, 14 PRs in 5 months
Contributions summary:Kejsi contributed to the `sktime` library, which focuses on machine learning for time series data. Their work included refactoring the `TSFreshClassifier`, updating the documentation for the `predict_quantiles` method, and implementing and adding tests for `predict_quantiles` and `predict_intervals` in the `theta` forecasters, alongside a fix for the `Prophet` forecaster to address issues with exogenous variables. The user also extended the functionality of the `pmdarima` adapter to include `predict_quantiles`.
forecastingtime-series-analysistime-series-regressiondata-sciencedeep-learning
kejsitake/sktime
Aug 2021 - Oct 2022
A unified framework for machine learning with time series
Contributions:106 pushes, 23 branches in 1 year 2 months
deep-learningtime-seriesmachine-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.