Website:
Job details:
Position Overview: Acutix3D is seeking a Computer Vision Engineer to build the "brain" of our autonomous metrology systems. You will develop the high-throughput semantic regression networks and 3D topological mapping algorithms that live directly on our camera interface. You will leverage the NVIDIA Jetson Orin NX to process complex multi-vector mapping in under 12 seconds, decoupling metrology from massive hardware and shifting it into agile edge compute.
Key Responsibilities:
- Algorithm Development: Develop and optimize computer vision algorithms for our symmetric quad-projection system to effectively neutralize shadows and specular glare on micro-components (0201 metric).
- Edge Compute Optimization: Deploy and optimize deep learning models on the NVIDIA Jetson Orin NX (utilizing its 100 TOPS and Dual NVDLA v2.0 accelerators) for strictly bounded <12s cycle processing times.
- Warp Correction: Implement software-driven dynamic coordinate recalculation to isolate true solder height by referencing the board (Z_ref(x,y) = ax+by+c), bypassing the need for expensive mechanical Z-tracking.
- Synthetic Initialization: Develop zero-shot synthetic initialization pipelines to allow for immediate algorithm stabilization without the need for multiple manual scrap runs.
- Continuous Learning: Contribute to our compounding SaaS features, including continuous self-learning adaptation and predictive maintenance analytics.
Qualifications:
- Strong academic and practical background in 3D Computer Vision, computational photography, or structured light (multi-angle fringe projection experience is highly desirable).
- Expertise in edge AI optimization using NVIDIA ecosystems (CUDA, TensorRT, DeepStream) and deploying models on resource-constrained edge devices.
- Proficiency in C++ and Python, with deep experience in AI/ML frameworks (PyTorch, TensorFlow).
- Solid mathematical foundation in 3D geometry, point cloud processing, and spatial transformations.
- Experience in the manufacturing or metrology space (defect detection, anomaly detection, topological mapping) is highly preferred.
Click on Apply to know more.
