Alex T.

Machine Learning Engineer

Alex har en Ph.D. i maskinlæring, og har mer enn 7 års yrkeserfaring innen ML og informatikk.

Han har jobbet som maskinlæringsingeniør i selskaper som Bosch og Samsung med brukssaker innen datasyn og dyp læring.

Hovedekspertise
  • NumPy
    NumPy 7 år
  • Python
    Python 7 år
  • C++
    C++ 3 år
Andre kunnskaper
  • TensorFlow
    TensorFlow 4 år
  • Pandas
    Pandas 4 år
  • Caffe
    Caffe 3 år
Alex
Alex T.
Kom i gang

Utvalgt opplevelse

Arbeidserfaring

  • Senior Machine Learning Engineer

    Virtual Control - 2 years

    • Development of a computer vision solution for shipments tracking in the supply chain.
    • Development of a human pose estimation algorithm based on deep learning technologies.
    • Development of human detection, one-camera tracking, multi-camera tracking, re-identification algorithm.
    • Increased accuracy of initial open source solution by 26% and inference time was decreased by 50%.
    • Designed architecture of the computer vision solution.
    • Improved speed of the system from 1 frame-per-second to 10 frames-per-second on 4 CPUs(without GPU usage) without loss in accuracy.
    • Led a team of the three ML and Software engineers in the project.

    Teknologier:

    • Teknologier:
    • NumPy NumPy
    • Python Python
    • Caffe Caffe
    • C++ C++
    • Kotlin Kotlin
    • OpenCV OpenCV
    • TensorFlow TensorFlow
    • Openvino Openvino
    • Kubeflow Kubeflow
  • Senior Software Engineer

    Occipital - 10 months

    • Development of ROS and ROS2 drivers for RGB-D camera.

    Teknologier:

    • Teknologier:
    • C++ C++
  • Senior Machine Learning Engineer

    Samsung R&D Institute - 1 year 3 months

    • Development of simultaneous localization and mapping (SLAM) algorithm for an autonomous vehicle.
    • Development of an object-detection algorithm for an autonomous vehicle based on deep learning technologies.
    • Development of traffic signs recognition algorithm for an autonomous vehicle based on deep learning technologies.
    • Improved accuracy of the initial open source solution by 5% by fusion of camera and lidar data.
    • Improved accuracy of the initial open source solution by 20% and inference speed was decreased by 30% by redesign of Neural Network architecture and improved training strategy.

    Teknologier:

    • Teknologier:
    • NumPy NumPy
    • Python Python
    • Caffe Caffe
    • C++ C++
    • Keras Keras
    • OpenCV OpenCV
    • TensorFlow TensorFlow
  • Machine Learning Engineer

    PrivatBank - 7 months

    • Development of a computer vision solution for an automatic line of people detection and counting.

    Teknologier:

    • Teknologier:
    • NumPy NumPy
    • Python Python
    • Caffe Caffe
    • XGBoost XGBoost
  • Software Engineer

    LightIt - 10 months

    • Development of a web-application with data analytics tools.

    Teknologier:

    • Teknologier:
    • Python Python
  • Mathematical Engineer

    JSC Zaporizhstal - 1 year 9 months

    • Development of a model of gas recirculation in sintering furnace and its implementation in the controller.
    • Development of a system for information gathering about all expenses of resources in a plant.

    Teknologier:

    • Teknologier:
    • Python Python
    • SQL SQL
    • MSSQL MSSQL
    • C++ C++

Utdannelse

  • Doctor Of PhilosophyMachine learning

    Taras Shevchenko National University of Kiev · 2018 - 2022

  • MSc.Machine Learning

    Taras Shevchenko National University of Kiev · 2012 - 2014

  • BSc.Automatisering og dataintegrerte teknologier

    Zaporizhia State Engineering Academy · 2008 - 2012

Finn din neste utvikler innen dager, ikke måneder

Vi kan hjelpe deg å levere produktet ditt fortere med en erfaren, ekstern utvikler. Fra 369 kr/h. Betal bare om du er fornøyd med din første uke.

I løpet av en kort 25-minutters samtale ønsker vi å:

  • Forstå dine utviklingsbehov
  • Forklare prosessen vår der vi matcher deg med kvalifiserte, evaluerte utviklere fra vårt nettverk
  • Dele de neste stegene for å finne riktig match, ofte på mindre enn en uke

Ikke sikker på hvor du begynner?

La oss ta en prat

Den første utvikleren begynner innen få dager. Ingen aggresive salgsargumenter.