Camarillo, California · hace 2 meses
Diseñar, prototipar e implementar algoritmos de procesamiento de imagen y visión por computadora, incluidos modelos CNN y SNN, para pipelines de cámaras EO/IR, e integrarlos con los equipos de firmware y FPGA en hardware en tiempo real.
SAAZ is seeking an exceptional Senior Computer Vision Algorithm & Software Engineer to architect and implement advanced processing pipelines for next-generation electro-optical and infrared (EO/IR) imaging systems. This multidisciplinary role bridges theoretical research and high-performance software engineering, requiring hands-on expertise in image processing, computer vision, convolutional neural networks (CNNs), or spiking neural networks (SNNs) rather than traditional application software development alone. Operating with a high degree of autonomy, the successful candidate will hold a graduate degree—preferably a PhD—and possess the independent drive to conceptualize, design, and deploy sophisticated algorithms from scratch without direct supervision. Working closely with Firmware, FPGA, Systems, and Product Engineering teams, you will drive algorithmic innovation for advanced camera products deployed in aerospace, defense, and commercial imaging, guiding developments from concept through hardware-software integration and production. Key Responsibilities: • Algorithm Architecture & Conceptualization: Design, prototype, and refine advanced image processing and computer vision algorithms—including traditional image enhancement, noise reduction, and modern deep learning models (CNNs/SNNs)—tailored for EO/IR sensor architectures. • Autonomous End-to-End Implementation: Independently translate mathematical models and theoretical concepts into high-performance, maintainable software implementations without needing direct step-by-step supervision. • Cross-Functional System Integration: Collaborate closely with Firmware, FPGA, and Systems Engineering teams to optimize, port, and validatealgorithms on real-time target hardware and embedded camera processing platforms. • EO/IR Pipeline Optimization: Develop and tune edge-detection, feature extraction, non-uniformity correction (NUC), dynamic range expansion, object detection and tracking algorithms specialized for complex electro-optical and infrared environments. • Research & Feasibility Trade Studies: Conduct independent trade studies, literature reviews, and rapid prototyping to evaluate novel machine learning and spiking neural network (SNN) approaches for low-power or bandwidth-constrained imaging systems. • Verification & Testing Pipelines: Build robust simulation environments, ground-truth dataset collection methodologies, and automated testing frameworks to evaluate algorithm accuracy, latency, and performance edge cases. • Technical Documentation & Mentorship: Document mathematical formulations, algorithmic trade-offs, and software architectures to support production handover, system qualification, and intellectual property development. Minimum / Required Qualifications: • Education & Experience: Master’s degree in Electrical Engineering, Computer Science, Applied Mathematics, Optical Engineering, or a closely related field with 5+ years of hands-on algorithmic experience, or a Ph.D. with 2+ years of relevant research/industry experience. • Domain Expertise: Proven hands-on track record developing and deploying core image processing, computer vision, or neural network models (CNNs or SNNs). Pure application software development without signal processing or computer vision experience will not be considered. • EO/IR Pipeline Knowledge: Direct familiarity with physical imaging concepts and sensor data pipelines, such as non-uniformity correction (NUC), bad pixel replacement, dynamic range compression, or thermal/optical noise reduction. • Core Technical Stack: Proficiency in C/C++ and Python for rapid prototyping, algorithm implementation, and performance benchmarking. • Mathematical Foundations: Strong foundation in linear algebra, multi-variable calculus, spatial/frequency-domain filtering, and statistical signal processing. • Autonomous Execution: Demonstrated ability to drive projects independently from literature review and mathematical formulation through to functional code without daily supervision. Preferred Qualifications: • Advanced Degree: Ph.D. focusing on Computer Vision, Spiking Neural Networks (SNNs), Neuromorphic Computing, or Infrared Image Processing. • Hardware-Aware Software Optimization: Experience adapting heavy algorithmic models or neural networks for resource-constrained platforms, such as embedded GPUs (NVIDIA Jetson), FPGAs, or specialized DSP architectures. • Advanced Neural Architectures: Hands-on research or deployment experience with Spiking Neural Networks (SNNs), event-based neuromorphic sensors, or ultra-low-latency event processing for edge execution. • Specialized EO/IR Algorithms: Background in real-time object detection/tracking, multi-sensor data fusion (EO/IR registration), or high dynamic range (HDR) image reconstruction. • Simulation & Frameworks: Expertise with deep learning and vision frameworks (e.g., PyTorch, OpenCV, TensorRT, LibTorch) alongside customized C++ execution pipelines. Success Metrics: • Algorithms execute within defined latency, framerate, and real-time processing constraints. • CPU, GPU, and embedded memory utilization remain within target resource budgets. • Algorithms achieve specified precision, recall, and detection/tracking accuracy metrics across target EO/IR datasets. • Code passes performance, memory leak, and static analysis checks without critical warnings or errors. • Compliance with project coding, testing, and documentation standards. • Unit, integration, and ground-truth simulation tests successfully completed. • Seamless pipeline integration and data throughput achieved across cross-functional interfaces (Firmware, FPGA, host APIs). • Algorithmic robustness validated under adverse real-world conditions (e.g., thermal drift, low contrast, high dynamic range, atmospheric noise).
Publicada el 13 ago 2026
