Nhat Nguyen

Software Engineer at VNDCREDIT
  • Claim this Profile
Contact Information
us****@****om
(386) 825-5501
Location
Hanoi, Hanoi, Vietnam, VN

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

  • Second prize at the National Student Physics Olympiad in 2018
    Vietnam Physical Society
    Apr, 2018
    - Sep, 2024

Experience

  • VNDCREDIT
    • Vietnam
    • Financial Services
    • 1 - 100 Employee
    • Software Engineer
      • Dec 2020 - Present

      Developing AI models and backend AI: Smart-camera, Face search, eKYC Smart camera • Trained, tuned and serving AI models for object detection and classification purposes. • Built a complete pipeline that solves intelligent traffic and face search problems based on deepstream SDK. This pipeline can handle up to 6 camera streams while ensuring real-time capability. Especially the face search system searches in real time on a database of up to 5M faces. • Customize pipeline’s input source with Gstreamer so it has the ability to add and remove during runtime. • Custom video management system (VMS) relies on sdk metadata provided by NetworkOptix to display pipeline results. • Built an intermediate video analytics (VA) to receive pipeline results via kafka and then process and send to VMS using socket protocol. Face search • Optimized and serving face detection and face feature extraction models with the ability to handle multiple requests and batch processing. • Implement approximation search algorithms to increase search speed by 100 times and reduce memory resources by 20 times. • Built a face search system on a database of 47M faces. The system only takes less than 0.1 seconds to return the list of results using approximation search and multi-processing algorithms. • Built a video face analysis system. The system can analyze how many people appear in the video, how often and when each person appears in the video. Performance can be up to 160 FPS when a 2-hour video only takes about 45 seconds to process. eKYC System • Trained, tuned and serving text detection model, text recognition model, face matching model and liveness model for eKYC system. • Embedded the id card detection model down to the mobile device. • Built a distributed logging system using filebeat, logstash, elasticsearch, kibana. • Built a load balancer block to serve the ekyc service cluster using nginx. Show less

  • G Group
    • United Kingdom
    • IT Services and IT Consulting
    • 1 - 100 Employee
    • Software Engineer
      • Jun 2019 - Nov 2020

      Built AI models to analyze images and videos of social network users - Trained and optimized the gender classification model, the age prediction model, the beauty evaluation model to group users according to the following criteria. - Fine-tuning model of video classification for newspaper video classification problem. - Develop a service to check for duplicate videos from the input queue. Built AI models to analyze images and videos of social network users - Trained and optimized the gender classification model, the age prediction model, the beauty evaluation model to group users according to the following criteria. - Fine-tuning model of video classification for newspaper video classification problem. - Develop a service to check for duplicate videos from the input queue.

Education

  • Hanoi University of Science and Technology
    2016 - 2021

Community

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