Qualcomm
Senior Computer Vision Algorithm Engineer – Hardware Co‑Optimization (Edge AI/SoC)
- Compensation
- Not disclosed
- Location
- Hsinchu City, Taiwan
- Work setup
- Onsite
- Seniority
- Senior
- Employment
- Not specified
- Experience
- 1+ years
- Category
- Hardware
- Posted
Skills
Must have
- Strong background in computer vision, especially segmentation (classical and deep learning–based).
- Solid understanding of SoC architecture, including compute accelerators, memory hierarchy, and hardware/software interfaces.
- Hands‑on experience with modern deep learning architectures: Transformers and efficient transformer variants FlashAttention or similar attention‑optimization techniques Mamba or other state‑space mode
- Applicants: Qualcomm is an equal opportunity employer. If you are an individual with a disability and need an accommodation during the application/hiring process, rest assured that Qualcomm is committ
- Qualcomm expects its employees to abide by all applicable policies and procedures, including but not limited to security and other requirements regarding protection of Company confidential information
- To all Staffing and Recruiting Agencies: Our Careers Site is only for individuals seeking a job at Qualcomm. Staffing and recruiting agencies and individuals being represented by an agency are not aut
- If you would like more information about this role, please contact Qualcomm Careers.
Nice to have
- Algorithm–hardware co‑design for edge AI
- Quantization, pruning, distillation, or other model‑efficiency techniques
- Real‑time CV systems or embedded AI deployment
- Performance profiling tools and hardware simulators
Full description
##
Company:
Qualcomm Semiconductor Limited
Job Area:
Engineering Group, Engineering Group > Machine Learning Engineering
General Summary:
We’re seeking a multidisciplinary engineer with strong expertise in computer vision algorithms and SoC/hardware architecture. This role is ideal for someone who thrives at the intersection of cutting‑edge AI models and efficient hardware implementation. You’ll design segmentation algorithms, optimize modern deep learning architectures, and help shape the hardware that runs them.
Responsibilities
- Develop and optimize computer vision algorithms, with emphasis on segmentation and real‑time performance.
- Implement and tune state‑of‑the‑art efficient AI architectures, including Transformers, Mamba‑style sequence models, and attention‑optimization techniques such as FlashAttention.
- Collaborate with hardware and SoC teams to co‑design accelerators and pipelines for CV and ML workloads.
- Profile and optimize model performance across memory, compute, and bandwidth constraints.
- Contribute to system‑level architecture decisions for next‑generation CV/AI products.
- Work closely with firmware and software teams to deploy models on embedded or edge platforms.
##
Qualifications
- Strong background in computer vision, especially segmentation (classical and deep learning–based).
- Solid understanding of SoC architecture, including compute accelerators, memory hierarchy, and hardware/software interfaces.
- Hands‑on experience with modern deep learning architectures:
- Transformers and efficient transformer variants
- FlashAttention or similar attention‑optimization techniques
- Mamba or other state‑space models
- SOTA efficient models (e.g., MobileNet‑style, ConvNeXt, lightweight segmentation networks)
- Proficiency in Python and C/C++, plus experience with ML/CV frameworks (PyTorch, TensorFlow, ONNX, OpenCV).
- Familiarity with hardware modeling, FPGA prototyping, or RTL design is a plus.
- Strong ability to work across algorithm, software, and hardware boundaries.
Preferred Experience
- Algorithm–hardware co‑design for edge AI
- Quantization, pruning, distillation, or other model‑efficiency techniques
- Real‑time CV systems or embedded AI deployment