Ilya Zhenin

Senior AI Developer at Umojo
  • Claim this Profile
Contact Information
us****@****om
(386) 825-5501
Location
Moscow, Moscow, Russia, RU
Languages
  • Russian Native or bilingual proficiency
  • English Professional working proficiency
Skills

Topline Score

Topline score feature will be out soon.

Bio

Generated by
Topline AI

You need to have a working account to view this content.
Join now
You need to have a working account to view this content.
Join now

Credentials

  • 9.01x: Light, Sight, Spike: The Neuroscience of Vision
    edX
    Jan, 2015
    - Nov, 2024
  • 8.01x: Physics I
    edX
    Jan, 2014
    - Nov, 2024
  • Machine Learning
    Coursera
    Jul, 2013
    - Nov, 2024
  • Introduction to Astronomy
    Coursera
    Jan, 2013
    - Nov, 2024
  • edX Honor Code Certificate for Light, Spike, & Sight: The Neuroscience of Vision
    edX

Experience

  • Umojo
    • United States
    • Information Technology & Services
    • 1 - 100 Employee
    • Senior AI Developer
      • Mar 2022 - Present

      Developing Computer Vision production pipeline software in C++: Ingesting incoming videostreams, pre-processing, sending them to inference, colleting and distributing results. System features distributed computations(over multiple GPUs and over multiple servers), high-safety and resilence, coming back after failures, restoring state, reconnecting to lost services, etc. Stack: - Triton Inference Server (C++) - Gstreamer (writing pipeline, stream discovery and different features in C API) - C++ (gcc on linux), multi-threaded program with thread pool for inference and with safe synchronization between different threads(mutexes, conditional variables). OOP, RAII. - GRPC server(code generation from proto and RPC calls implementation in C++) for remote interface. I.e. interface makes gprc calls to server to update itself with information - AMQP client(for RabbitMQ) interface with libuv in C++ - Boost libs (command line args parsing and hopefully URI scheme when 1.81 comes out) - JSON lib, logging(spdlog), date::format lib, yaml-cpp - Model conversion to torchscript - OpenCV - Python On ML side: developing algorithms for segmentation processing Show less

  • Mars Vision
    • Software Development
    • Senior ML Engineer
      • Jun 2019 - Mar 2022

      Developing video games visual assistance software for visually impaired people. Navigation in video games, aiming assitance, sound clues and developing accessible interface. - C++ desktop Windows client, featuring screen capture, gaming overlay, neural network processing(detection, tracking, depth estimation), aiming assitance, QT interface tuned for accessibility - Training neural networks, detection, mono depth estimation, traicking algorithms developments(using keypoints and proximity metrics) Stack: - Pytorch, Tensorflow - Python - C++ - QT - OpenCV Show less

  • Huawei
    • China
    • Telecommunications
    • 700 & Above Employee
    • Senior Developer Team Lead
      • Apr 2016 - Jun 2019

      Teamlead developer of Huawei MRC team, working in research and development in the field of image processing and artificial intelligence. Main expertise: - Organizing team work to achieve project goals - Building top neural networks solutions for Semantic Segmentation and Instance-Aware Semantic Segmentation for photo / video segmentation - Object Detection - Object tracking in videostream - Depth reconstruction on image stereo pair by classical and neural network approaches - Depth reconstruction on single image - Neural Network convergence analysis and loss landscape exploration - Data augmentations - Expertise in different approaches and architectures: ResNet, DenseNet, Pyramidal Net, ASPP, SqueezeNet, AlexNet, Inception, VGG, MobileNet-v1-2, YOLO-v1-v3, Mask-RCNN, U-Net and so on Python, C++, Caffe, TensorFlow Show less

  • ICEBERG Sports Analytics
    • Canada
    • Spectator Sports
    • 1 - 100 Employee
    • Computer Vision Research and Development Engineer
      • Sep 2015 - Feb 2016

      Building computer vision system for detection,tracking and identification of hockey players. Building computer vision system for detection,tracking and identification of hockey players.

  • NTMR OOO
    • Appliances, Electrical, and Electronics Manufacturing
    • Computer Vision Research and Development Engineer
      • Sep 2013 - Sep 2015

      System and mathematical models development based on machine learning and computer vision algorithms for object detection and recognition given three-dimensional radio-wave image of a person(s), implemented in security systems(radio scaners). System and mathematical models development based on machine learning and computer vision algorithms for object detection and recognition given three-dimensional radio-wave image of a person(s), implemented in security systems(radio scaners).

  • Paragon Software
    • United States
    • IT Services and IT Consulting
    • 1 - 100 Employee
    • SW Programmer
      • 2013 - 2013

      Parsers development for dictionary databases. Development of morphological parsers. Building a publishing system. Parsers development for dictionary databases. Development of morphological parsers. Building a publishing system.

Education

  • НИЯУ МИФИ
    Applied Mathematics
    2010 - 2013
  • Coursera
  • МГУПИ

Community

You need to have a working account to view this content. Click here to join now