вчера

Machine Learning Compiler Engineer

ориентир по рынку
вакансия зп не указана
в среднем 149 218 ₽
Загрузи резюме, чтобы видеть мэтчи с вакансиейЗагрузи резюме, чтобы видеть мэтч с вакой

подготовься к отклику

ai-инструменты

Чтобы адаптировать резюме под вакансию или составить сопроводительное письмо, загрузи резюме

описание

We’re looking for compiler engineers to join a new ML Compilers Team at Jane Street.

Machine learning is a critical pillar of our global business. Our ever-evolving trading environment serves as a unique, rapid-feedback platform for ML experimentation, allowing us to incorporate new ideas with relatively little friction. Our ML team is full of people with a shared love for the craft of software engineering, and for designing APIs and systems that are delightful to use.

ML compilers let us quickly turn research ideas and novel architectures into high-throughput training jobs and low-latency inference. To get the most out of our hardware resources and enable rapid model development and deployment, we are working to understand and improve the current ML compiler technologies and explore where it makes sense to build our own solutions. In this role, you would shape the development of our compiler stack.

If you’ve never thought about a career in finance, you’re in good company. Many of us were in the same position before working here. If you have a curious mind and a passion for solving interesting problems, we have a feeling you’ll fit right in.

задачи

Задач нет

требования

The ideal candidate has worked on an ML compiler in the past, but non-ML compiler experience is also valuable. Experience with any part of the stack is fine, be it high-level ML frameworks (PyTorch, Jax, TensorFlow), graph compilers (Inductor, XLA, TVM) or kernel DSLs (Triton, Helion, cuTile). You should understand how all those pieces fit together even if you are unfamiliar with specific parts.

Some other things you might be excited about are:

  • Optimisations in ML compilers: algebraic rewriting, materialise vs recompute tradeoffs and fusion, equality saturation, scheduling languages, loop-level optimisation, autotuning
  • ML compiler frontends: graph capture, symbolic and dynamic shapes, control flow, automatic differentiation
  • GPU kernel programming and architecture: Triton, cuTile, CuTeDSL, CUDA, compute and memory hierarchies, programming with tensor cores
  • Compiler concepts: static analysis, intermediate representations like SSA and CFGs, instruction selection
  • ML inference: CUDA Graphs, KV caching, batching
  • ML training: Tensor/Sequence/Expert/Pipeline Parallelism, FSDP, NCCL, activation checkpointing
  • GPU debugging and optimisation: CUDA-GDB, Nsight Systems, Nsight Compute
  • Formal methods: proof assistants, automated theorem provers, agentic proof systems
  • Programming languages and DSL design: embedded DSLs, staging and metaprogramming, type systems, functional programming
  • Fluency in English is required

If you're a recruiting agency and want to partner with us, please reach out to agency-partnerships@janestreet.com.

условия

Условий нет

Если просят выйти из iCloud, прислать код из SMS, запустить или установить что-то, перевести деньги — не соглашайся: это мошенничество.

Если просят выйти из iCloud, прислать код из SMS, запустить или установить что-то, перевести деньги — не соглашайся: это мошенничество.