About The Role

AI Vision Processors For Edge Applications

Our solutions make cameras smarter by extracting valuable data from high-resolution video streams.

Job Description

Join us to build Classical CV and NNISP acceleration pipelines on CVflow, and improve system performance and efficiency.
We are looking for engineers passionate about embedded systems and computer vision — fresh graduates and experienced candidates are welcome.
___________________________________________________________________________

About the Team

We are the CV Accelerator team within the DSP department,focused on delivering high-performance edge-side computer vision processing on Ambarella’s proprietary CVflow™ architecture.

Our work targets real-world applications requiring low latency, low power, and real-time performance at the edge.

___________________________________________________________________________

What You’ll Do

Port and optimize classical (non-neural-network) computer vision algorithms onto CVflow (e.g., image processing, filtering, optical flow, radar/LiDAR point cloud processing)
Perform NNISP model porting for neural network–based ISP pipelines
Develop efficient implementations on:

NVP (Neural Vector Processor)

GVP (General Vector Processor)

Optimize compute performance and memory efficiency
Ensure real-time performance on edge hardware platforms
Leverage modern AI-assisted coding tools to improve productivity
Collaborate with cross-functional teams (algorithm / system / hardware)

___________________________________________________________________________

Why Join Us

Work on edge AI / computer vision acceleration hardware (CVflow)
Gain experience in both classical CV and NNISP pipelines
Tackle real-world performance-critical systems
Grow into an expert in edge computer vision and hardware-aware optimization
Be part of a team using modern, high-productivity development workflows

Preferred Qualifications

Familiarity with embedded systems (ARM-based)
Understanding of multithreading (Linux or RTOS)
Experience in debugging and performance optimization
Background in image processing or computer vision is a strong plus
Familiarity with Python or deep learning frameworks (e.g., PyTorch) is a plus, especially for NNISP-related development

___________________________________________________________________________