Image Algorithm Engineer

About Company

Established in 2022, XG Tech is pioneering the evolution of smart vehicles by advancing automotive digital transformation from distributed computing systems to centralized, cross-domain platforms. Their specialization lies in intelligent cockpit technology—considered the next major competitive differentiator—while integrating sophisticated driving assistance solutions. XG Tech envisions future vehicles as mobile living spaces, aligning with trends positioning cars as a "third living space."

Role Summary and Responsibilities

The Image Algorithm Engineer will be responsible for creating and refining computer vision solutions tailored for automotive functionalities, including driver monitoring systems (DMS), around view monitoring (AVM), vehicle security such as Sentry, and privacy features like de-identification. Key tasks involve developing lightweight vision algorithms optimized for embedded automotive hardware, emphasizing constraints such as limited compute resources, low latency, and high reliability.

The role extends to managing full lifecycle vision data pipelines from data acquisition, annotation protocol creation, data cleansing, augmentation, to model training, deployment, and iterative improvements. There is also an emphasis on innovative integration, combining internal cockpit perception with external vehicle environment sensing and decision-making systems.

Additional focus areas include exploring applications powered by vision-language models (VLM), large language models (LLM), and AI Agents to facilitate intelligent multimodal interactions and on-device real-time AI. Responsibilities cover deploying and optimizing such models on automotive edge computing chips, including compression, acceleration, and heterogeneous computing techniques for efficient, low-power inference.

Desired Qualifications and Expertise

• Minimum of 3 years' experience in developing computer vision algorithms, with at least 1 year specializing in automotive cockpit applications like DMS, OMS, or IMS.
• Proficiency in object detection, keypoint detection, segmentation, tracking techniques, and deep learning frameworks such as PyTorch and TensorFlow.
• Direct experience with model training, optimization, and deployment targeting embedded platforms.
• Familiarity with model conversion and acceleration toolchains such as ONNX, TensorRT, TFLite, and neural processing unit (NPU) toolkits.
• Experience integrating in-cabin perception systems with vehicle decision-making or autonomous driving functionalities.
• Knowledge and experience with VLM/LLM, including multimodal modeling, fine-tuning, and optimizing on-device inference.
• Competence in developing AI Agent systems involving task planning, memory management, tool invocation, streaming, and multimodal contextual perception and decision-making.