Welcome
A Mid-Level Machine Learning Engineer with intensive skills in Machine Learning and Computer Vision. He had almost 4 years of experience on different projects that use ML and CV to solve real-world problems.
His daily job is to build Machine Learning pipelines, and models that improve the accuracy of various processes. Also, researching, designing, implementing, and deploying scalable computer vision and machine learning solutions.
Frequently used
Experience
VDSMART .JSC, Vietnam
Head of Machine Learning
Nov 2019 - Present
- * Optimized Edge device (Jetson NX) with Deepstream NVIDIA to run Face/Activity Recognition real-time up to 4 cameras, 2M pixels
- * Researched Activity recognition model: Sleep, Hand raising with GCN architecture with 95% accuracy for School application
- * Designed the Face/Activity Recognition pipeline for serving AI with Edge device and Server with Triton Server Nvidia for over 5000 people
- * Built AI Server for serving 1000 requests/min with Uvicorn
- * Designed AI Server with microservice architecture and integrate to K8S
- * Built Searching Face service for 1 million Faces in database with 0.03s/request using Faiss Facebook
- * Converted model from Tensorflow, Pytorch to TensorRT Jetson NX to speed up
- * Deployed AI with Amazon SageMaker, Amazon Web Services (AWS), REST APIs and Docker
- * Experienced with Hadoop, MapReduce, Spark Expert knowledge developing production level ML products
- * Experienced with A/B Testing ML models in production using Amazon SageMaker
- * Managed the function's process activities, work commitments and performance, day-to-day oversight to AI team with Trello
- * Designed take-home coding assignments for interview candidates, and conduct onsite interview
- * Built Face Recognition System and integrate to Barrier Gate for Face timekeeper with 1000 people in company
- * Implemented Face Quality Assessment model to remove low quality Face
- * Researched U-net model for satellite image (sentinel image) segmentation to detect deforestation in Western Highlands - Vietnam
- * Suggested technical AI ideas for project manager
Awards
1st Prize, AIoT Developer InnoWorks
Year 2019
This project won the first prize in AIOT developer innoworks 2019 hosted by ADVANTECH CO., LTD which applied AI - Image segmentation in Spectrum Images captured by Drone to detect Rice yellow dwarf pathogens
1st Prize, Panasonic IoT Innovation Idea Contest
Year 2018
This project won the first prize in Panasonic IoT Innovation Idea Contest 2018 which applied Food recognition and Food recommendation technology for refrigerator
