Machine Learning Compiler Engineer - Apple Neural Engine (Front-End) at Apple

Sunnyvale, California, United States -

Full Time

Start Date

Immediate

Expiry Date

20 Feb, 26

Salary

0.0

Posted On

22 Nov, 25

Experience

5 year(s) or above

Remote Job

Yes

Telecommute

Yes

Sponsor Visa

Skills

Machine Learning, Compiler Engineering, C++, MLIR, Graph Optimization, Debugging, Neural Networks, SoCs, Deep Learning, JIT Compilation, Multi-threading, Collaboration, Communication, Performance Optimization, Error Diagnostics, Visualization

Industry

Computers and Electronics Manufacturing

Description

At Apple, we're on the cutting edge of delivering transformative experiences through Artificial Intelligence. If you're passionate about pushing the boundaries of AI and hardware optimization, we want you to join our team! As a Machine Learning Compiler Engineer (Front-End) on our Apple Neural Engine (ANE) team, you'll work to bring high-performance, low-power AI solutions to life on iconic Apple products like the Vision Pro, iPhone, iPad, Mac, and more. This is a dynamic opportunity to work with us in a creative, collaborative environment while developing groundbreaking technologies that will shape the future of computing! Are you ready to help us deliver the next groundbreaking Apple products? DESCRIPTION As a Machine Learning Compiler Engineer (Front-End), you will be empowered to: • Implement or extend MLIR dialects tailored to capture semantics of ANE operations with a focus on ML types and operators • Build conversion and lowering passes from higher-level dialects like Metal Performance Shaders graph or CoreML, including canonicalization passes to make higher-level dialects more ANE-friendly • Build validation passes to support placement on ANE, including validation of computational graphs for atomic placement of groups of operations • Build passes and tools to aid debugging of functionality (ex: numerical accuracy) and performance (ex: fusion passes) • Improve compiler efficiency and model asset size with infrastructure and passes to reduce constant duplication by tracking mutation of weight tensors • Develop tools for test coverage, error diagnostics, and visualization in the MLIR pipeline • Collaborate with higher layer dialects for integration of new features and with compiler backend and runtime engineers to align with lower-level IRs and code to ensure alignment with HW constraints and performance goals MINIMUM QUALIFICATIONS Bachelor's degree in Computer Science, Computer Engineering, or a related field with 5 years of relevant experience Interest/background in MLIR framework & tooling Interest/background in compiler frontend & IR design, including canonicalization and graph optimization techniques Solid debugging and code navigation of complex compiler pipelines Strong experience in C++ or similar object-oriented programming language PREFERRED QUALIFICATIONS Master's/Ph.D. degree in Computer Science, Computer Engineering, or a related field 10 years of relevant experience Demonstrated ability to ship high-quality production software - Experience optimizing compilers for distributed, parallel, or heterogeneous execution environments, with a solid understanding of shared memory, synchronization, and multi-threading techniques Expertise in neural network inference on specialized SoCs or GPUs, and knowledge of deep learning frameworks and tools Familiarity with Just-in-Time (JIT) compilation and dynamic optimization techniques for real-time code execution Proven track record in mentoring and coaching engineers, with an interest in taking on increasing responsibilities and contributing to the team's development Strong collaboration and communication skills across teams (ex: across compiler layers and runtime framework teams)

Responsibilities

As a Machine Learning Compiler Engineer (Front-End), you will implement or extend MLIR dialects for the Apple Neural Engine. You will also build conversion passes, validation passes, and tools for debugging and performance optimization.