Our client is SI
Responsibilities:
· Develop and optimize real-time computer vision applications for video analytics and edge AI solutions.
· Design and implement data pipelines to handle high-volume, real-time video streams.
· Build scalable, high-performance video processing pipelines with integration into cloud services or on-prem infrastructure.
· Collaborate with machine learning engineers to deploy deep learning models for object detection, segmentation, tracking, and classification on live video feeds.
· Ensure system performance, scalability, and robustness through optimization and efficient resource utilization.
· Participate in the full software development lifecycle, from requirements gathering to deployment and maintenance.
· Stay updated on the latest advancements in computer vision, machine learning, and deep learning technologies, incorporating them into the project roadmap.
Requirements:
· 3+ years of experience in computer vision or related fields.
· Expertise in NVIDIA DeepStream SDK for real-time video analytics and AI-based applications.
· Hands-on experience with Apache Kafka for building real-time data pipelines.
· Extensive experience in AWS
· Proficiency in OpenCV, GStreamer for image and video processing tasks.
· Strong experience with C++, Python, and libraries/frameworks such as
· TensorFlow, PyTorch, or TensorRT for deploying AI models.
· Understanding of video compression standards (H.264, H.265) and streaming protocols (RTSP, RTP, etc.).
· Familiarity with Docker and Kubernetes for deploying scalable microservices in cloud environments.
· Experience with Edge AI hardware like NVIDIA Jetson or similar platforms.
· Solid understanding of parallel computing and GPU acceleration techniques (CUDA).
· Strong problem-solving skills and ability to work in a fast-paced, collaborative environment.
Chin, SF
sf.chin[at]jondavidsongroup.com
JonDavidson Pte Ltd
Co. Reg. No. 200709068N. Licence No. 22S1412
Registration Number: R21100141
"Personal data collected will be used for recruitment purposes only"
