AI Performance Engineer (Competitive & Network Analysis)

About Qualcomm

Qualcomm is expanding its presence in Riyadh as part of Saudi Arabia's Vision 2030 digital transformation, investing in advanced computing and data centre capabilities to power AI, cloud, and connectivity at scale. This role offers a chance to contribute to the development of data centre operations and Cloud AI technology in the region.

Role Overview

We are seeking an AI Performance Engineer to join a collaborative team working on the entire product lifecycle from research to deployment. The position involves optimizing AI model inference, working with advanced generative AI algorithms, and collaborating across multiple teams to enhance AI workload performance on current and future hardware platforms.

Main Responsibilities

• Optimize and deploy AI models using frameworks like PyTorch and ONNX for efficient inference.
• Analyze and improve performance of large language, vision-language, and diffusion models considering throughput and latency.
• Map next-generation AI workloads to hardware architectures to maximize efficiency.
• Collaborate closely with customers and internal teams including compiler, firmware, and platform experts.
• Diagnose and resolve complex issues affecting performance or stability.
• Develop engineering solutions for continuous AI workload performance insights and improvements.
• Design and implement efficient, high-level kernels using tools like Triton.

Required Qualifications and Skills

• Proven experience in building and optimizing language models, preferably in production using PyTorch and ONNX.
• Strong understanding of transformer models, attention mechanisms, and their performance trade-offs.
• Experience with workload distribution techniques such as sharding and parallelism.
• Advanced Python programming capabilities.
• Active engagement with latest inference optimization methods.
• Knowledge of computer architecture, machine learning accelerators, in-memory processing, and distributed systems.
• Excellent communication and problem-solving skills; ability to thrive in a fast-paced, collaborative environment.
• Master's degree in Computer Science, Machine Learning, Computer Engineering or Electrical Engineering.

Preferred (Bonus) Qualifications

• Background in neural network operators, linear algebra, and mathematical libraries.
• Familiarity with machine learning compilers.
• Experience in balancing model accuracy and evaluation methodologies.
• Knowledge of torch.compile or torchDynamo utilities.
• PhD in Computer Science, Computer Engineering, or Machine Learning.

Compensation and Benefits

• Competitive salary including housing and transportation allowances.
• Stock options (RSUs) and performance bonuses.
• 16 weeks fully paid maternity leave and 6 weeks fully paid paternity leave.
• Employee stock purchase program.
• Child education allowance.
• Relocation and immigration assistance if necessary.
• Life and medical insurance coverage.
• Wellness reimbursement for health and fitness memberships.

Minimum Eligibility

• Bachelor's degree plus 4+ years relevant systems engineering experience, or
• Master's degree plus 3+ years relevant systems engineering experience, or
• PhD plus 2+ years relevant systems engineering experience.
• Equivalent experience considered if candidate demonstrates required competencies.

Additional Information

Qualcomm is committed to equal opportunity employment and accessibility, offering accommodations for candidates with disabilities. Employees are expected to comply with company policies regarding confidentiality and security. Recruitment agencies are not authorized to submit applications on behalf of candidates.