You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
description: 'Working on edge AI solutions, real-time video processing, and high-performance systems for traffic monitoring and analysis.',
22
22
achievements: [
23
-
'Developed a high-performance storage solution for traffic metadata analysis on edge devices, achieving 5x performance improvement over traditional NoSQL storage systems',
24
-
'Created traffic congestion detection algorithm capable of classifying 7 levels of congestion, achieving 95% accuracy at Bach Lieu Bridge for real-time traffic management',
25
-
'Improved License Plate reading pipeline efficiency using weighted random algorithm, increasing plate reading ability by 30-40%',
26
-
'Developed custom NAL buffer for RTSP streaming, ensuring stable stream quality and low-latency performance under varying network conditions',
27
-
'Built asynchronous video pipeline handling up to 4 FHD RTSP streams for AI Hub, optimizing real-time video processing and resource utilization',
28
-
'Engineered real-time live view system for edge AI camera with ~10ms latency and minimal resource usage (1% CPU, 50MB RAM on ARM Cortex-A53)',
29
-
'Optimized and deployed computer vision models on Novatek SoCs, achieving real-time 30fps processing for full AI pipeline (vehicle detection, LPR, OCR)',
23
+
'Developed a high-performance storage solution for traffic metadata analysis on edge devices, achieving a 5× performance improvement over traditional NoSQL storage systems.',
24
+
'Implemented a circular buffer–based recording pipeline for 5MP (2448×2048) encoded video, achieving a 1.5 GB RAM reduction at 15 FPS.',
25
+
'Created an algorithm to analyze traffic data and detect congestion, capable of classifying seven levels of traffic congestion. Successfully implemented the algorithm to efficiently detect traffic jams at Bach Lieu Bridge up to ~95% of times, improving real-time traffic management and response.',
26
+
'Improved license plate recognition (LPR) pipeline efficiency using weighted random sampling which increased plate reading ability by 30-40%',
27
+
'Developed a custom NAL buffer for RTSP streaming, ensuring low-latency and stable video quality under varying network conditions.',
28
+
'Built an asynchronous video pipeline supporting up to 4 FHD RTSP streams, optimizing real-time processing and resource utilization.',
29
+
'Engineered a real-time live view system for an edge AI camera, achieving ~10 ms latency with minimal resource usage (~1% CPU, 50 MB RAM) on ARM Cortex-A53.',
30
+
'Optimized and deployed computer vision models on Novatek SoCs, achieving real-time 30 FPS across the full AI pipeline (vehicle detection → LPR → OCR) while maintaining low resource consumption.',
31
+
'Designed and implemented an IPC library enabling multiple processes to share a single web server, aggregating H.264/H.265 streams via a centralized socket and distributing them over WebSocket to concurrent clients.',
description: 'Get lost in a remote corner, look for the living using an old radio, stare at something blankly while listening to music, try your best to survive—you may have never imagined such scenes but it seems as if they occur every day. In the game, you will be in danger at any time and you need to be courageous and smart enough to survive!',
0 commit comments