Talk to sales Try for Free
Taras Vozniuk

Taras Vozniuk

Research Software Engineer, Computer Vision

Followers of Taras Vozniuk409 followers
location of Taras VozniukSeattle, Washington, United States

Connect with Taras Vozniuk to Send Message

Connect

Connect with Taras Vozniuk to Send Message

Connect

  • Timeline

  • About me

    Staff Machine Learning Engineer at Cruise, ex-Meta, PhD

  • Education

    • Kyiv National Taras Shevchenko University

      2006 - 2010
      Bachelor's degree Applied Mathematics 5.0

      With honors (Red diploma)Theme of my bachelor work:"Semi-supervised learning for document classification"

    • Kyiv National Taras Shevchenko University

      2012 - 2016
      Doctor of Philosophy (PhD) Computer Science

      Topic : "The use of semantic-syntactic tensor model of natural language for analysis of coreferential relations in text"Thesis is dedicated to improvement of systems for finding coreferential connections in the text using the tensor model of natural language. A large text corpus was analyzed and a six-dimensional tensor of typical syntactic structures was built. New efficient algorithms for constructing the control space of syntactic structures was developed to improve the tensor model… Show more Topic : "The use of semantic-syntactic tensor model of natural language for analysis of coreferential relations in text"Thesis is dedicated to improvement of systems for finding coreferential connections in the text using the tensor model of natural language. A large text corpus was analyzed and a six-dimensional tensor of typical syntactic structures was built. New efficient algorithms for constructing the control space of syntactic structures was developed to improve the tensor model. This made possible to save more information about the semantic and syntactic relationships in tensors of less dimension. Show less

    • Taras Shevchenko National University of Kyiv

      2010 - 2012
      Master's degree Computer Science 5.0

      With honors(red diploma)Theme of my masters work : "Semantic methods for solving problems of anaphora" (Computational Linguistics)

  • Experience

    • Viewdle

      Jul 2010 - Sept 2012
      Research Software Engineer, Computer Vision

      I was working on Face and feature detection, Face recognition/clustering/tracking.One of my biggest accomplishments here was improving Face grouping by 15% on average set by introducing some structural rules to our face clustering algorithm.I wrote different tuning applications to improve quality of face recognition by automatic choosing best parameters of image color normalization and face grouping thresholds.Tools and technologies : C++, STL, Boost, Java, SVN/Git, Teamcity, Jenkins, JIRA, Redmine Show less

    • Motorola Mobility (a Google Company)

      Oct 2012 - Dec 2015
      Senior Research Engineer, Computer Vision

      I was a member of ATAP group led by former DARPA director Regina Dugan, working on Highlight Reel application, which creates short videos with best moments from photos and videos on your phone.- Implemented detection and training part of deformable parts model for facial landmark detection (by article of Deva Ramanan, CVPR-2012). This method uses HoG features combined with quadratic feature location penalty, uses SVM for training, and dynamic programming for fast detection. I wrote effective code in C++ with ARM NEON optimization for mobile and developed heuristics to speed up this method.- Optimized PCA calculation over DCT features using ARM NEON and preallocated pool of objects in memory.- Implemented and integrated effective cache for faceprint search in SQLite to speed up face recognition pipeline.- Ongoing work on integrating internal projects after API changes, automatic accuracy tests on Jenkins, making app deliveries. Show less

    • Innit

      Dec 2015 - Dec 2016
      Senior Software Engineer (Computer Vision)

      Led raw food and packages recognition efforts (Computer Vision, Pattern Recognition)Used state of the art deep learning techniques to develop raw food recognition. Trained AlexNet, VGG-19 and ResNet classifiersDeveloped local feature based framework for package and logo detection and recognition.

    • Meta

      Dec 2016 - Aug 2023

      Support a team of 10 engineers, including 8 ML engineers and 2 ML infrastructure engineers, focusing on expanding the use of personalization and ML across the pillar. TL on the team of 6-8 eng, working on improving commenting in News Feed by suggesting short text, sticker and emoji replies. Responsible for ML and ML infra part of the project, partnering with frontend team and PM. Working in NewsFeed Discovery team. Training ranking models on petabytes of data to find most relevant stories (Ranking, Big Data, Machine Learning)

      • Software Engineering Manager, Machine Learning

        Jan 2022 - Aug 2023

      • Staff Software Engineer, Machine Learning

        Jul 2019 - Dec 2022

      • Senior Software Engineer, Machine Learning

        Dec 2016 - Jun 2019

    • Cruise

      Oct 2023 - now
      Staff Applied Scientist TLM, Perception

      On the Computer Vision Foundation team, built and deployed ML models for object detection and 3D position estimation from fisheye cameras in the ultra-near field, while supporting two ML engineers.

  • Licenses & Certifications