Senior Software Engineer – EDGE AI Development
QuEST Global
Job Requirements
Key ResponsibilitiesDesign and Development:Design and implement AI/ML-based applications tailored for embedded and edge hardware.Develop end-to-end pipelines for model training, conversion, and deployment.Customize neural network architectures for edge-specific applications such as object detection, classification, and segmentation.Model Optimization and Deployment:Port and optimize AI models to meet performance and memory constraints on edge platforms.Apply quantization (e.g., INT8), pruning, and layer fusion techniques to improve model efficiency.Convert models between various formats such as ONNX, TensorRT, TVM, TFLite, and DRP-AI.Performance Tuning and Profiling:Analyze model bottlenecks and tune for latency, throughput, and power efficiency.Run inference performance profiling on hardware targets and iterate for improvements.Testing and Debugging:Validate model accuracy and performance post-optimization.Debug and troubleshoot model behavior discrepancies across frameworks and devices.Documentation and Research:Maintain documentation for all model lifecycle stages and optimization steps.Stay updated with latest AI compiler advancements and deployment trends in edge AI.
Work Experience
Must Have:Bachelor's/Master’s degree in Computer Science, Electronics, or AI-related field.6+ years in AI/ML model development with experience in real-world applications.Proficient in Python, C++ and deep learning libraries (TensorFlow, PyTorch, Keras).Solid understanding of CNNs, FCNs, and their applications in computer vision.Practical knowledge of model optimization workflows (quantization, pruning, etc.).Experience with ONNX, TVM, TensorRT, DRP-AI, TFLite, OpenCV, etc.Experience with deployment on edge devices like Jetson, RZ/V2H, or STM32.Strong understanding of constraints (compute, memory, power) in edge environments.Good to Have:Exposure to embedded Linux or RTOS environments.Familiarity with low-level model debugging, calibration tools, and inference engines.Experience with Continuous Integration tools such as Git, Jenkins, or similar.Understanding of hardware accelerators (GPU, NPU, DRP-AI, etc.).
Key ResponsibilitiesDesign and Development:Design and implement AI/ML-based applications tailored for embedded and edge hardware.Develop end-to-end pipelines for model training, conversion, and deployment.Customize neural network architectures for edge-specific applications such as object detection, classification, and segmentation.Model Optimization and Deployment:Port and optimize AI models to meet performance and memory constraints on edge platforms.Apply quantization (e.g., INT8), pruning, and layer fusion techniques to improve model efficiency.Convert models between various formats such as ONNX, TensorRT, TVM, TFLite, and DRP-AI.Performance Tuning and Profiling:Analyze model bottlenecks and tune for latency, throughput, and power efficiency.Run inference performance profiling on hardware targets and iterate for improvements.Testing and Debugging:Validate model accuracy and performance post-optimization.Debug and troubleshoot model behavior discrepancies across frameworks and devices.Documentation and Research:Maintain documentation for all model lifecycle stages and optimization steps.Stay updated with latest AI compiler advancements and deployment trends in edge AI.
Work Experience
Must Have:Bachelor's/Master’s degree in Computer Science, Electronics, or AI-related field.6+ years in AI/ML model development with experience in real-world applications.Proficient in Python, C++ and deep learning libraries (TensorFlow, PyTorch, Keras).Solid understanding of CNNs, FCNs, and their applications in computer vision.Practical knowledge of model optimization workflows (quantization, pruning, etc.).Experience with ONNX, TVM, TensorRT, DRP-AI, TFLite, OpenCV, etc.Experience with deployment on edge devices like Jetson, RZ/V2H, or STM32.Strong understanding of constraints (compute, memory, power) in edge environments.Good to Have:Exposure to embedded Linux or RTOS environments.Familiarity with low-level model debugging, calibration tools, and inference engines.Experience with Continuous Integration tools such as Git, Jenkins, or similar.Understanding of hardware accelerators (GPU, NPU, DRP-AI, etc.).
Por favor confirme su dirección de correo electrónico: Send Email