Computer Vision & Edge AI Engineer

Purpose of the Job:Develop, train, adapt, and optimize deep learning models for computer vision applications running on edge AI devices. The role covers the complete model lifecycle, including developing new model architectures, training and retraining models, modifying pretrained networks, and optimizing inference performance for target hardware while balancing accuracy, computational complexity, and real-time performance.Responsibilities and Duties:Develop and implement computer vision models for object detection, image segmentation, classification, and other vision-based tasks.Research and evaluate state-of-the-art architectures and techniques to address application-specific requirements.Prepare datasets and develop preprocessing, augmentation, and training pipelines.Train, fine-tune, and retrain models using PyTorch or similar deep learning frameworks.Modify pretrained model architectures, layers, and operators to improve accuracy, efficiency, and hardware compatibility.Optimize models using quantization, pruning, precision, layer fusion, and other model compression techniques.Convert and optimize models using ONNX, TensorRT, and hardware-specific AI compilers and SDKs.Deploy and validate models on edge AI platforms equipped with GPUs, NPUs, or dedicated AI accelerators.Profile and benchmark model accuracy, inference latency, FPS, memory consumption, and hardware utilization.Troubleshoot model conversion, compilation, numerical accuracy, and runtime performance issues.Collaborate with embedded software engineers to integrate models into production applications and inference pipelines.Maintain reproducible model development, training, evaluation, optimization, and deployment workflows.Education Bachelor's or master's degree in computer engineering, Computer Science, Artificial Intelligence, Electronics Engineering, or related fields.Strong academic foundation in deep learning, computer vision, and neural network architectures.Experience

• 0-1 Practical experience developing, training, and evaluating computer vision models.
• Hands-on experience with PyTorch, pretrained model fine-tuning, transfer learning, and neural network architecture modification.
• Experience preparing datasets, implementing data augmentation, and evaluating model performance.
• Experience with ONNX export, computational graph optimization, and model conversion.
• Practical experience optimizing and deploying deep learning models on edge AI devices.
• Experience with model quantization, mixed-precision inference, and hardware-aware optimization.
• Experience with TensorRT or hardware-specific AI compilers and SDKs is preferred.
• Automotive, robotics, or other real-time computer vision experience is an advantage.Skills and Abilities
• Strong programming skills in Python/C++ and proficiency with PyTorch, OpenCV, and NumPy.
• Strong understanding of deep learning architectures, training strategies, loss functions, and model evaluation metrics.
• Ability to design new models and modify existing network architectures to meet functional and computational requirements.
• Knowledge of model compression, quantization, pruning, and inference acceleration techniques.
• Ability to analyze computational graphs and resolve unsupported operators and model conversion issues.
• Experience profiling inference latency, throughput, memory usage, and hardware utilization.
• Strong analytical, debugging, and problem-solving skills.
• Ability to balance model accuracy, computational efficiency, and real-time performance constraints.
• Good collaboration and technical communication skills.