Natalia Pavlovskaia

NATALIA PAVLOVSKAIA Deep Learning Engineer Berlin, Germany ¯ https://www.linkedin.com/in/yanebo/ ‡ https://github.com/ne-bo drivendata https://www.drivendata.org/users/Natalia.Pavlovskaia/ kaggle https://www.kaggle.com/nebo86 DATA SCIENCE EXPERIENCE PROJECTS Deep Learning Engineer 3 audatic.ai 11.2022 – 05.2025 Berlin, Germany I’m working on speech enhancement for hearing aid, use Tensorflow. • Target Speaker extraction by the voice sample/direction • Speech Enhancement (denoising) • Models Compression (knowledge distillation, NAS, quantization) Machine Learning Specialist Giant.ai 04.2020 – 07.2022 Larnaca, Cyprus Drivendata competition – 10-th place out of 1217 Time series analysis. Sustainable Industry: Rinse Over Run competition – bronze medal 3 Kaggle Images retrieval. Google Landmark Retrieval Challenge 2019 competition – bronze medal 3 Kaggle Images retrieval. Google Landmark Retrieval Challenge 2018 I worked on the humanoid robot trained using reinforcement learning to do tasks such as pick and place. • Perception: autoencoders, contrastive learning, 3D object pose estimation. Robot achieved >90% success rate. EDUCATION • Sensors fusion: made the system robust to the loosing the data from 1-2 Higher School of Economics & Skoltech out of 4 different sensors. Master, Statistical learning theory - Moscow, Russia • Dynamics: created a model of the robot’s dynamics which improved the success rate in simulation by 5%. • Behavioral cloning: trained policies for different tasks with success rates above 90%. • Engineering: made the real robot 3.3 times faster. Research Scientist (Computer Vision) Speech Technology Center 07.2019 – 04.2020 Saint-Petersburg, Russia Face recognition system can be fooled by the presentation of a printed photo or replayed video. So I worked on the passive face antispoofing task. • data collection (academic data, scraping, crowd-sourcing) • deep learning models implementation and training (using Pytorch) Specialist (5 years), Astronomy, Graduated with Honors. Saint-Petersburg State University - Saint-Petersburg, Russia NON DATA SCIENCE WORKING EXPERIENCE Software Testing Engineer • moving and testing the trained models in C++ compatible format 2011 – 2017 • creation the working demo, including the front-end (JS) • Deutsche Bank, 2015 – 2017 My main contribution was the domain adaptation part which brought an improvement of 10% HTER. • Yandex, 2013 – 2015 Deep Learning Engineer Online The most successful project I’ve completed as a freelancer is an asthma condition prediction for the asthma patients on the next day. The F1-score is above 0.7. https://erj.ersjournals.com/content/56/suppl_64/155 Computer Vision Intern Junior Researcher The Central Astronomical Observatory of the RAS at Pulkovo 2008 – 2010 Neuromation 07.2018 – 08.2018 • OOO ROLIS, 2011 – 2013 I wrote tests in Java. Managed a team of 3 persons. Upwork 08.2018 – ongoing Saint-Petersburg, Russia Saint-Petersburg, Russia I’ve developed scalable classification system of retail products. This system allows automatically recognize the products SKUs on a supermarket’s shelf. I’ve used images retrieval approach. Classification accuracy on unseen classes was at 96% level which is on par on existing production solutions from other companies. Saint-Petersburg, Russia