AI Engineer

Stack

Responsibilities

• ▹Design, build and maintain real-time 3D spatial processing pipelines using sensor fusion (LiDAR, camera feeds, spatial telemetry) for point cloud filtering, segmentation and 3D layout analysis
• ▹Develop computer vision models and post-processing algorithms to extract structural features, recognize building geometry and generate precise 2D/3D floorplans from raw visual and spatial data
• ▹Translate research findings, algorithmic prototypes and modern AI/ML/LLM concepts into high-performance, maintainable production code in Python, taking direct ownership of core product implementations
• ▹Architect, deploy and manage multi-model inference pipelines on AWS SageMaker and NVIDIA Triton Inference Server, ensuring low-latency processing and reliable high-throughput serving
• ▹Build end-to-end data and MLOps pipelines (synthetic data generation, active annotation, dataset versioning, CI/CD, real-time telemetry) to continuously evaluate, deploy and monitor model performance, latency and spatial accuracy
• ▹Work with software engineering, research and product management teams to transition prototype features into scalable, market-ready releases

Requirements

• ▹Production-level mastery of Python alongside working knowledge of C++ or Swift, emphasizing clean code, modular design and execution speed
• ▹Hands-on experience with OpenCV, Open3D or PCL (Point Cloud Library) for point cloud filtering, spatial segmentation, feature extraction and 2D/3D coordinate transformations
• ▹Deep experience with PyTorch or TensorFlow, alongside proficiency in Scikit-learn for traditional machine learning and statistical data analysis
• ▹Proven experience building low-latency serving infrastructure using NVIDIA Triton Inference Server and managing end-to-end model workflows on AWS SageMaker
• ▹Familiarity with model quantization, pruning and target compilers (e.g. ONNX Runtime, TensorRT) to hit production latency targets
• ▹Solid foundation in linear algebra, 3D geometry, coordinate systems and multi-sensor fusion, and awareness of modern LLM/multimodal applications

Nice to have

• ▹Wireless domain knowledge: basic understanding of RF environment simulation, indoor spatial coverage modeling or wireless network planning
• ▹Experience optimizing or running vision models on iOS devices (CoreML, ARKit, Metal)

Soft skills

About the company

Ekahau offers software and hardware products that let IT professionals manage the full Wi-Fi lifecycle: design, site survey and troubleshooting. Ookla is part of Accenture.