Chenhsi Sung

senior Computer Vision and ML Research Engineer/Data Scientist at mindtrace.ai
  • Claim this Profile
Contact Information
us****@****om
(386) 825-5501
Location
London, England, United Kingdom, UK
Languages
  • English -
  • Chinese (Traditional) -
  • German -
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

  • Deep Learning Specialization
    Coursera
    Jul, 2022
    - Nov, 2024
  • Sequence Models
    Coursera
    Jul, 2022
    - Nov, 2024
  • 機器學習技法 (Machine Learning Techniques)
    Coursera
    Jul, 2022
    - Nov, 2024
  • Convolutional Neural Networks
    Coursera
    Dec, 2020
    - Nov, 2024
  • Structuring Machine Learning Projects
    Coursera
    Nov, 2020
    - Nov, 2024
  • Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
    courser
    Oct, 2020
    - Nov, 2024
  • Neural Networks and Deep Learning
    Coursera
    Aug, 2020
    - Nov, 2024
  • 機器學習基石下 (Machine Learning Foundations)---Algorithmic Foundations
    Coursera
    Jun, 2020
    - Nov, 2024
  • 機器學習基石上 (Machine Learning Foundations)---Mathematical Foundations
    Coursera
    May, 2020
    - Nov, 2024

Experience

  • mindtrace.ai
    • United Kingdom
    • Software Development
    • 1 - 100 Employee
    • senior Computer Vision and ML Research Engineer/Data Scientist
      • Aug 2021 - Present

      • Spearheaded 3D object classification initiatives and drove the development of the main production pipeline. • Pioneered an unsupervised learning strategy for data clustering, achieving in a remarkable 90% accuracy rate in benchmark assessments. • Enhanced pipeline maintenance and optimized functionalities for a 25% increase in overall efficiency. • Integrated 3D instance segmentation into the pipeline, significantly advancing the capabilities of the current unsupervised stack. • Implemented GNN solutions on a comprehensive dataset, as a result an impressive 20% boost in accuracy, surpassing previous benchmarks. Show less

  • Self-employed
    • Computer Vision and consultant
      • Jan 2021 - Present

      • Conducted comprehensive code maintenance, astute refactorization, and detailed profiling, leading to substantial enhancements in algorithm efficiency. • Conducted comprehensive code maintenance, astute refactorization, and detailed profiling, leading to substantial enhancements in algorithm efficiency.

  • Helmholtz Munich
    • Germany
    • Research Services
    • 700 & Above Employee
    • Computer Vision and ML Engineer
      • Jan 2021 - Aug 2021

      • Engineered machine learning and deep learning applications specializing in 2D/3D object detection and classification. • Conducted comprehensive code maintenance, astute refactorization, and detailed profiling, leading to substantial enhancements in algorithm efficiency. • Mastered training of advanced models, specializing in detecting 3D point clouds in order to remove artifacts within construction sites. • Engineered machine learning and deep learning applications specializing in 2D/3D object detection and classification. • Conducted comprehensive code maintenance, astute refactorization, and detailed profiling, leading to substantial enhancements in algorithm efficiency. • Mastered training of advanced models, specializing in detecting 3D point clouds in order to remove artifacts within construction sites.

  • Contilio
    • United Kingdom
    • IT Services and IT Consulting
    • 1 - 100 Employee
    • Computer Vision & AI Engineer
      • Jan 2019 - Dec 2020

      • Constructed an AABB tree structure, innovatively optimizing adjacent 3D objects connectivity. • Effectively employed advanced outlier noise removal techniques and achieved a remarkable 30% reduction in noisy data. • Accelerated point cloud classification efficiency by 9% through launching a multi-processing pipeline. • Expertly trained CNN models, executing object detection tasks with an exceptional accuracy of 83%. • Refined the ray-tracing algorithm, strategically escalating the quantity of mesh coverages for improved rendering precision. Show less

  • IDEAi LLC
    • Taipei City, Taiwan
    • R&D engineer
      • Feb 2017 - Sep 2018

      Key tasks and achievements: 1. Implemented an AR open source SDK for tracking and object recognition using 3D Computer Vision: - Launched prototypes for multiple platforms. - Increased ARToolKit detection range from 1 direction to 4 directions. - Improved Unreal ARToolKit plugin to enable a detection up to 20 markers. 2. Improved a scalable, large and multi-threaded video software platform, optimizing for best performance and stability: - Decreased video transmission time by 6% by improving server-side implementation. - Improved our product user experience by eliminating repetitive input and functionality. Show less

  • Intel Corporation
    • United States
    • Semiconductor Manufacturing
    • 700 & Above Employee
    • Intel R&D internship
      • Jan 2015 - Mar 2015

      1.Researched and evaluated state of the art computer vision and ML techniques 2.Implemented several prototypes for comparing classification and segmentation performance 3.Conducted final presentation and discussion with the entire R&D group 1.Researched and evaluated state of the art computer vision and ML techniques 2.Implemented several prototypes for comparing classification and segmentation performance 3.Conducted final presentation and discussion with the entire R&D group

Education

  • The University of Edinburgh
    Master's degree, Computer Science
    2015 - 2016
  • National Tsing Hua University
    exchange program, Computer Science
    2014 - 2015
  • National Taiwan University
    Bachelor’s Degree, Horticulture and Landscape Architecture
    2009 - 2013

Community

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