Makhmud Shaban

Makhmud Shaban

CV/DL Engineer Intern

Followers of Makhmud Shaban472 followers
location of Makhmud ShabanLimassol, Cyprus

Connect with Makhmud Shaban to Send Message

Connect

Connect with Makhmud Shaban to Send Message

Connect
  • Timeline

  • About me

    Founding Computer Vision Research Engineer @ Nim

  • Education

    • Higher School of Economics

      2017 - 2021
      Bachelor's degree Applied Mathematics and Information Science
  • Experience

    • Yandex

      Jul 2019 - Nov 2019
      CV/DL Engineer Intern

      Worked on transferring image feature extraction pipeline to a newer framework - TensorFlow.

    • Cognitive Dynamic Systems Lab, MIPT

      Jul 2020 - Jul 2021
      Deep Learning Researcher

      Conducted various experiments and participated in research in the field of Visual Question Answering.Successfully trained SOTA model with modification in scene parsing - using Vector Symbolic Architectures to enhance the ability of the model to generalize on VQA task extensions. Co-authored a paper on the following topicSuccessfully adapted the following approach to Visual Dialog and co-authored a paper on the results of such adaptation.

    • Prisma Labs, Inc.

      Oct 2021 - Dec 2023
      Research And Development Engineer

      Worked on a broad spectrum of features, including Style Transfer, Image Generation and general Computer Vision algorithms.Lensa:1. Neck shadow - applying realistic shadow that makes the visual separation of face and neck more apparent- Designed, developed and refined mobile-friendly neck shadow algorithm, which made it to production and is available on Lensa app.2. "Magic Avatars" for Lensa- Participated in the development of "Magic Avatars" - a viral neural network portrait generation feature, which enabled Lensa to get to the top of App Store/Play Store and attracted millions of users worldwide.3. Charm AI - Conducted various experiments to improve the quality of face beautification solution.Prisma: 1. Oil painting style - Conducted extensive experiments that improved the quality and got rid of artifacts in oil painting stylization, on-device model successfully integrated.Improved general speed of conducting experiments by reimplementing augmentations, which led to 4x speedup and removed bottleneck in preparation of images. Show less

    • Higher School of Economics

      Nov 2021 - Dec 2021
      Guest Instructor

      - Conducted Research Seminar which consisted of 4 classes- Co-developed a seminar plan for the students

    • Nim

      Feb 2024 - now
      Founding Computer Vision Research Engineer
  • Licenses & Certifications

    • Certificate of Proficiency in English

      Cambridge Assessment English
      Jun 2017
      View certificate certificate
  • Honors & Awards

    • Awarded to Makhmud Shaban
      Kaggle Competitions Expert -