Key Responsibilities
* Embedded Deployment: Collaborate with FPGA hardware, embedded software, and data science teams to deploy AI solutions directly onto satellites and other constrained Edge AI devices.
* Algorithm Conversion: Implement high-performance solutions by transferring and optimizing algorithms initially created in Python into robust C/C++ codebases.
* Infrastructure Development: Research, recommend, and implement new hardware and software solutions to improve the company’s overall AI infrastructure.
* Performance Optimization: Ensure AI models are highly optimized for efficiency, especially when utilizing hardware accelerators like GPUs and NPUs.
* Team Support: Provide computational and deployment support to the Remote Sensing and Data Science teams.
Required Qualifications
* Masters/PhD in Computer Science, Engineering, or a related technical field.
* Fluency in English.
* More than 2 years of professional experience in Embedded Software Development.
* Programming fluency in C, C++, and Python.
* Proficiency with Linux environments and collaborative development using GitHub.
* Experience with hardware acceleration technologies, including Graphics Processing Units (GPUs) and Neural Processing Units (NPUs).
* Expertise in ML/Deep Learning deployment frameworks, such as TensorFlow Lite, ONNX runtime, or PyTorch Edge.
* Working knowledge of MLOps principles for training/evaluation pipelines and automated model delivery/monitoring.
Critical Bonus Skills
* Experience with AMD Versal AI engines and Vitis Model Composer (Kernel development, data flow optimization, model quantization/pruning, Vitis IDE, and performance analysis are a plus).
* Experience deploying models via web services, dashboards, and APIs (e.G., FastAPI, Flask, gRPC) and using cloud services/containerization (GCP, AWS, Azure, Docker, Kubernetes).
* Familiarity with High-Performance Computing (HPC) and job scheduling systems like SLURM.
Nice to Have
* Familiarity with containerization on constrained systems (e.G., Singularity, microcontainers).
* Knowledge of data compression techniques for in-orbit data handling.
* Prior experience in the aerospace or remote sensing industries.
#J-18808-Ljbffr