(Sr./Staff) Software Engineer, Edge AI - Toolchain

Role Overview

OMNIVISION is seeking experienced Senior or Staff Software Engineers to strengthen its Edge AI and Systems Software team. The group sits within the broader ASIC organization and focuses on designing and improving neural network solutions for the video processor roadmap, including support for the OAX8000, OAX4600, and OA7100 product lines.

The role is part of a compact team responsible for creating and advancing the toolchain that enables customers and partners to develop AI models or migrate existing ones so they run efficiently on the company’s Neural Processing Units (NPUs).

What You Will Do

• Define and build new toolchain capabilities for both existing and upcoming NPU architectures, in close collaboration with the inference engine team.
• Improve, sustain, and troubleshoot current toolchain functionality, while also adapting model network layers to achieve better NPU efficiency in partnership with application engineering teams and, at times, customers or partners.
• Continuously raise team productivity by introducing automation across workflows and development activities.

Technical Profile

• Strong hands-on Python development experience, particularly with class-based product design and established design patterns.
• Deep knowledge of compiler fundamentals, including development and debugging approaches, code generation, program analysis, and pipeline optimization.
• Effective communication and collaboration abilities, along with practical experience in estimating tasks, reporting progress, and using CI/CD and coordination tools such as Confluence, Git, and JIRA.
• Helpful but not mandatory: experience with deep learning compilers such as TVM, the PyTorch framework, Eclipse-based toolchain frontend development, and CPU/DSP/GPU/NPU architectures.

Requirements

• A Ph.D., MS, or BS in Computer Science or Computer Engineering.
• Minimum 5 years of professional software development experience.
• At least 3 years of that experience must be in compiler-related work.

Additional Information

This role supports the development of toolchains that help AI models perform optimally on NPUs. It is based in Singapore and is a full-time, onsite position.