Full Time Computer Vision Engineer Jobs, in PLUS Malaysia Petaling Jaya - Ricebowl

The incumbent of this role will be responsible to autonomously build, deploy, and scale fast, reliable, and maintainable Computer Vision products within a specialized domain, balancing high-accuracy model inference with real-time processing constraints, infrastructure efficiency, and cross-squad architectural alignment.

Key Responsibilities

• Deliver highly reliable Computer Vision features by writing clean, modular, and optimized pipelines that achieve target accuracy metrics, minimize false positives/negatives, and meet the "Definition of Done."
• Drive the squad’s technical growth through structured reviews of both software code and model architectures, ensuring all engineers are proficient in the "Universal Stack" and modern Computer Vision deployment frameworks.
• Design modular Computer Vision processing pipelines that easily integrate across squads via robust API contracts, making it simple for the TRE team to monitor frame-ingestion, queuing, and model outputs.
• Manage technical dependencies between upstream camera/video feeds and downstream engineering applications, coordinating integration testing to ensure smooth rollouts without ecosystem bottlenecks.
• Collaborate with UX Specialists to deliver low-latency visual data processing that powers high-performing, real-time user interfaces and intuitive tracking dashboards.
• Continuously optimize model inference and video pre/post-processing code to minimize latency and CPU/GPU strain, ensuring technical and hardware constraints never compromise the end-user experience.
• Engage closely with Product Owners to translate commercial roadmaps into technical execution strategies, ensuring engineering activities are accurately prioritized, sequenced, and sized against business goals.
• Identify and resolve technical debt within the Computer Vision codebase (such as memory leaks from video decoding or inefficient frame caching) to protect overall system health and prevent performance degradation.
• Partner closely with the Technology Reliability Engineering (TRE) team to ensure all CV models and ingestion workflows are documented for stability, instrumented for drift detection, and built for rapid recovery during field or stream failures.