PyTorch Operator - ML Engineer

PyTorch Operator - ML Engineer [$100-$160/hr] As referral partner, we are posting to seek experienced PyTorch experts who excel in extending and customizing the framework at the operator level. Ideal contributors are those who deeply understand PyTorch’s dispatch system, ATen, autograd mechanics, and C++ extension interfaces. These contractors bridge research concepts and high-performance implementation, producing clear, maintainable operator definitions that integrate seamlessly into existing codebases. Key Responsibilities • Design and implement new PyTorch operators and tensor functions in C++/ATen • Build and validate Python bindings with correct gradient propagation and test coverage • Create “golden” reference implementations in eager mode for correctness validation • Collaborate asynchronously with CUDA or systems engineers who handle low-level kernel optimization • Profile, benchmark, and report performance trends at the operator and graph level • Document assumptions, APIs, and performance metrics for reproducibility Ideal Qualifications • Deep understanding of PyTorch internals (TensorIterator, dispatcher, autograd engine) • Strong background in C++17+ and template metaprogramming within PyTorch’s ecosystem • Experience authoring or extending PyTorch custom ops or backends • Working knowledge of performance profiling tools and GPU/CPU interplay • Strong written communication and ability to deliver well-documented, self-contained modules • Prior open-source contributions to PyTorch, TorchInductor, Triton, or related projects are a plus More About the Opportunity • Ideal for contractors who enjoy building clean, high-performance abstractions in deep learning frameworks • Work is asynchronous, flexible, and outcome-oriented • Collaborate with CUDA optimization specialists to integrate and validate kernels • Projects may involve primitives used in state-of-the-art AI models and benchmarks Application Process • Share a concise summary of your experience with PyTorch internals and systems-level programming • Include links to open-source work, GitHub PRs, or sample operator implementations • Provide hourly rate, availability, and relevant technical background • Selected experts may complete a short, paid pilot module to demonstrate fit