Software Performance Engineer, Vision Products Group

About the role

As an engineer in this role you will help with ANE efficiency as well as improving overall visionOS responsiveness. You will work closely with Computer Vision and Foundation Model teams to diagnose performance bottlenecks and develop innovative solutions to optimize compute and memory footprint. Your responsibilities will include: Analyze, benchmark and identify performance bottlenecks for ML/Foundation models in Vision Pro Develop in-house performance solutions (such as analysis tools) to optimize inference stack & identify performance wins. Conduct deep-dive analysis to identify & reduce bottlenecks and measure inference improvements in terms of memory, compute and latency. Partner with feature and product teams to consult on algorithm and SW architectural design decisions to achieve visionOS performance goals Optimize SW and HW via tunable parameters & integrate research findings into product implementation

Responsibilities

• Designing and implementing computer vision and machine learning algorithm optimizations; CPU C++ implementations and python bridge to TensorFlow/Pytorch;
• Identify performance bottlenecks in software C++ code using advanced CPU profiling technologies and in ML networks.
• Analyzing performance issues and resolving performance regressions reported by Apple's users around the world.
• Developing microbenchmarks and performance validation tests to catch performance regressions in computer vision and machine learning algorithms to preserve the performance quality of Apple's products.
• Working on integration of Foundation Models (including LLMs) into Apple Vision Pro and making sure it meets power and performance targets.
• Working with Apple's customers to troubleshoot performance issues and suggest optimizations across a variety of use cases; and
• Working on analytical and simulation models to predict application, algorithm, OS and real time interactive graphics performance and power.

Requirements

• Familiarity with PyTorch, Tensorflow or CoreML
• Knowledge of AI/ML fundamentals including model evaluation and deployment/inference.
• Familiarity with optimizing model architectures for on-device inference
• Familiarity with Computer Vision, Transformers and LLM architectures.
• Exceptional written and verbal communication skills
• Minimum BS and 1+ years of relevant industry experience
• Familiarity with Operating Systems, embedding programming, parallel programming.
• Academic or industry experience using C/C++, Python or Swift
• Understanding of compilers, runtimes, system performance, system software engineering and ML architectures.
• Familiarity with optimizing inference latency, memory and compute.
• Familiarity with Gen AI coding tools. Demonstrated ability to use them for design, code generation, and debugging.
• Pay & Benefits
• Note: Apple benefit, compensation and employee stock programs are subject to eligibility requirements and other terms of the applicable plan or program.

Additional Information

The Vision Product Group is working on exciting new technologies. We are looking for a driven and dedicated performance engineer. This team is building the core foundational platform for some of Apple's most advanced technologies in spatial computing. As part of our creative organization, you will have a uniquely rewarding opportunity to craft future products that will delight and inspire millions of people every day.

Your Match

How well this role fits your profile.

Company Intel

What employees say

Worked at Apple? Share your experience