Orca AI
Orca AI
Orca AI dezvoltă sisteme de inteligență artificială și viziune computerizată pentru industriile maritimă, transporturilor și logisticii. Produsele sale SeaPod și FleetView sprijină navigația autonomă prin îmbunătățirea conștientizării situaționale a echipajului, automatizarea funcțiilor de supraveghere și furnizarea de informații operaționale în timp real. Compania se concentrează pe sprijinirea operatorilor de nave în luarea unor decizii de navigație mai sigure, îmbunătățirea eficienței și promovarea unor operațiuni maritime mai sustenabile.

Edge Computer Vision Engineer in Israel (Hybrid)

Develop and optimize real-time computer vision systems for Orca AI’s maritime safety technology. Improve inference speed, reliability, and observability on hardware deployed aboard vessels.

Descriere

  • Develop and optimize real-time computer vision pipelines processing live maritime video on edge systems
  • Convert research models into reliable, production-ready components deployed aboard vessels
  • Improve inference performance through quantization, pruning, and graph-level optimization
  • Profile and optimize the complete multi-camera pipeline, including video ingestion, preprocessing, inference, and postprocessing
  • Find and resolve performance constraints involving CPU, GPU, memory, and pipeline coordination
  • Evaluate and justify tradeoffs among latency, accuracy, stability, and resource usage
  • Design reliable video-to-output data and inference pipelines that deliver actionable information to crew
  • Create benchmarking and evaluation processes for end-to-end performance measurement and release decisions
  • Develop observability capabilities and debugging processes for production systems
  • Establish and maintain well-defined interfaces between research code and production software
  • Partner with research and backend teams to bring new models into production
  • Improve system efficiency and reliability within hardware and runtime limitations

Cerințe

  • At least five years of experience building production systems, including two or more years focused on computer vision or real-time video pipelines
  • Practical experience developing and operating computer vision or deep learning systems in production
  • Advanced modern C++ skills, including multithreading and real-time programming, plus strong Python experience
  • Experience developing for edge or embedded platforms such as NVIDIA Jetson
  • Strong knowledge of inference optimization technologies such as CUDA, TensorRT, ONNX, GStreamer, or DeepStream, along with CPU, GPU, memory, and latency constraints
  • A profiling-led approach to optimization and performance tuning
  • Ability to debug complex systems and assess behavior in noisy real-world environments
  • Experience building custom, high-performance data or inference pipelines is a strong advantage
  • Familiarity with multi-sensor fusion, such as combining camera and radar data, is a strong advantage
  • Experience deploying and operating machine learning models in production is a strong advantage
  • Experience with low-level optimization or C++ performance tuning is a strong advantage
  • Demonstrated experience optimizing model inference with TensorRT, ONNX Runtime, quantization, pruning, or comparable methods is a strong advantage

Locuri de muncă similare