AI & Computer Vision Research Engineer (R&D)

Petscast is pioneering the next generation of embodied AI outdoor safety wearables for pets. Unlike traditional reactive GPS trackers, Petscast bridges the fatal time gap in pet accidents by deploying real-time, on-device computer vision to identify toxic flora, dangerous wildlife, traffic threats, and terrain hazards before injuries occur.

We are seeking an AI & Computer Vision Research Engineer to lead applied research and algorithmic innovation in real-time edge computer vision. In this R&D role, you will investigate, design, and validate lightweight neural network architectures and novel object detection/segmentation techniques optimized for ultra-low-power embedded edge devices under dynamic outdoor environments.

• Applied Algorithm Research: Conduct original investigation into lightweight object detection, fine-grained plant/animal classification, and zero-shot/few-shot learning models to recognize obscure toxic flora and hazardous wildlife species.

• On-Device Model Optimization & Quantization: Research novel post-training quantization (PTQ), quantization-aware training (QAT), pruning, and knowledge distillation techniques to execute real-time inference on resource-constrained MCU/NPU hardware.

• Dynamic Outdoor Vision Robustness: Investigate solutions to technical uncertainties in edge computer vision, such as extreme motion blur, varying lux levels, lens distortion, and rapid camera translation on moving animals.

• Specialized Dataset & Synthetic Data Research: Design experimental methodologies for synthetic data generation (Generative AI/NeRFs/Diffusion) and automated dataset curation to solve long-tail distribution challenges in rare flora and fauna species.

• Research Benchmarking & Evaluation: Establish scientific testing frameworks to evaluate model accuracy, latency, power consumption, and memory footprint on embedded hardware platforms.

  Requirements & Qualifications

  • Education: Bachelor’s degree in Computer Science, Artificial Intelligence, Computer Engineering, Electrical Engineering, or a related STEM discipline.

  • Technical Expertise:

    • Proven experience in deep learning frameworks (PyTorch, TensorFlow) and CV libraries (OpenCV).

    • Hands-on research experience in model compression (TensorRT, CoreML, TFLite, ONNX Runtime, NNAPI).

    • Solid understanding of deep learning research literature in object detection (YOLO variants, MobileNet, EfficientDet, Edge-Transformers).

  • Research Mindset: Demonstrated ability to formulate hypotheses, design controlled experiments, analyze empirical data, and solve unstructured technical challenges.