In short
AI engineer developing and optimising computer vision algorithms for video analytics — detection, tracking, and event analysis for uses like intrusion, PPE, and falling-object detection and people counting. Day-to-day work is Python with OpenCV and PyTorch, training/tuning models and integrating them into production systems.
About the role
Salary: $4,800 – $7,000 per month
About the role
Develop and optimise computer vision algorithms for video analytics applications, transforming real-world visual data into reliable detection, tracking, and event analysis capabilities. You will contribute to production-ready solutions across applications such as intrusion detection, loitering detection, PPE detection, falling object detection, people counting, vehicle monitoring, and other abnormal event detection use cases.
Key responsibilities
Develop and optimise algorithms for object detection, classification, segmentation, object tracking, and event detection.
Build multi-frame video analysis logic using tracking, trajectory, direction, speed, region of interest, and temporal information.
Train, fine-tune, test, and evaluate computer vision models using real project datasets.
Analyse false positives and false negatives, improving model accuracy through model tuning, data improvement, tracking, and post-processing.
Develop reusable computer vision modules for detection, tracking, event analysis, and alarm triggering.
Work closely with software and deployment engineers to integrate algorithms into production systems.
What we're looking for
Bachelor's Degree in Computer Science, Computer Engineering, Electrical or Electronic Engineering, Artificial Intelligence, or a related field. Fresh graduates are welcome to apply.
Good programming skills in Python.
Hands-on experience with OpenCV and PyTorch or similar deep learning frameworks.
Experience with object detection models such as YOLO, RT-DETR, Faster R-CNN, or equivalent.
Knowledge of object tracking and video analytics.
Experience training, testing, or improving computer vision models.
Good understanding of image processing and multi-frame video analysis.
Ability to write clean, modular, and maintainable code.
Experience with ByteTrack, DeepSORT, BoT-SORT, ReID, or other tracking algorithms is preferred.
Experience with optical flow, background subtraction, segmentation, or pose estimation is preferred.
Familiarity with ONNX or TensorRT is preferred.
Experience with CCTV, security, smart city, industrial vision, factory automation, or traffic analytics applications is preferred.
Experience with VLM-based visual analysis or verification is an advantage.