Machine Learning Compiler Engineer

Overview

Jane Street is seeking compiler engineers to join its new Machine Learning Compilers Team based in London, United Kingdom. Machine learning serves as a critical pillar for the global business, providing a rapid-feedback platform for experimentation that allows the firm to incorporate new ideas. The team develops ML compilers to quickly translate research ideas and novel architectures into high-throughput training jobs and low-latency inference. Engineers in this role will work to understand and improve current ML compiler technologies, explore the creation of custom solutions, and shape the overall development of the compiler stack.

Who can apply

Ideal candidates have previous experience working on an ML compiler, though non-ML compiler experience is also valued. Applicants can have experience with any part of the stack, including high-level ML frameworks such as PyTorch, Jax, and TensorFlow, graph compilers like Inductor, XLA, and TVM, or kernel domain-specific languages including Triton, Helion, and cuTile. Candidates must understand how these different pieces fit together even if they are unfamiliar with specific parts. Fluency in English is required. Visa sponsorship may be offered for suitable candidates.