Edge Computer Vision Accelerator Engineer

Job Title

Edge Computer Vision Accelerator Engineer

Role Summary

Develop and optimize classical computer vision and NNISP (neural-network ISP) pipelines for Ambarella's CVflow architecture, focusing on low-latency, low-power edge performance.

Work on algorithm porting, vector-processor implementations, and system-level performance tuning within the CV Accelerator team in the DSP department.

Experience Level

Mid-level. The role is open to fresh graduates as well as candidates with relevant experience; no specific years of experience are stated.

Responsibilities

Implement and optimize vision and NNISP pipelines for edge hardware with an emphasis on performance and efficiency.

• Port and optimize classical (non‑NN) computer vision algorithms to CVflow (image processing, filtering, optical flow, point cloud processing).
• Port NNISP models for neural-network–based ISP pipelines.
• Develop efficient implementations for NVP (Neural Vector Processor) and GVP (General Vector Processor).
• Optimize compute performance, memory usage, and ensure real‑time operation on edge platforms.
• Use AI-assisted coding tools to improve development productivity.
• Collaborate with algorithm, system, and hardware teams to integrate and validate solutions.

Requirements

Key technical skills and experience expected or strongly preferred.

• Must-have: Familiarity with embedded systems (ARM-based) and experience with multithreading on Linux or RTOS.
• Must-have: Experience in debugging, profiling, and performance optimization for constrained hardware.
• Must-have: Practical ability to implement and optimize code for vector/SIMD-style processors (or equivalent low-level compute engines).
• Must-have: Experience ensuring real-time performance and working with memory- and compute-constrained platforms.
• Nice-to-have / Preferred: Background in image processing or computer vision.
• Nice-to-have / Preferred: Familiarity with Python or deep learning frameworks (e.g., PyTorch) for NNISP work.
• Nice-to-have / Preferred: Exposure to hardware accelerators, SIMD/vector processing, and AI-assisted coding tools.

Education Requirements

Not specified.

About the Company

Company: Ambarella

Headquarters: Santa Clara, California, USA

Ambarella is a leader in computer vision and video processing technology, providing advanced solutions that enhance the performance of video applications. Focused on quality and innovation, Ambarella develops products for diverse uses, including autonomous vehicles, surveillance systems, and smart cameras, aiming to deliver pristine imagery and efficient compression while minimizing power consumption.