Real-Time ML Systems Performance Engineer

Janestreet; Greater London; On-site; GBP 70,000 - 90,000; Full time

Job summary

Janestreet is seeking an engineer with a focus on low-level systems programming and optimisation to join our ML team in Greater London. The role involves enhancing the performance of machine learning models, ensuring efficient large-scale training and real-time inference.

Candidates should have strong knowledge in GPU technologies and an understanding of modern machine learning techniques. Debugging, optimisation skills, and fluency in English are essential for this position.

Qualifications

• Experience in low-level systems programming and optimisation.
• The experience and systems knowledge to debug training runs end to end.
• Intuitive understanding of CUDA memory hierarchy and GPU performance.

Responsibilities

• Optimise the performance of ML models for training and inference.
• Ensure efficient large-scale training in the trading environment.
• Improve systems-level performance and throughput analytics.

Skills

Understanding of modern ML techniques and toolsets

Low-level GPU knowledge

Debugging and optimisation experience

Library knowledge of Triton, CUTLASS, CUB

Intuition about latency and throughput in CUDA

Background in Infiniband, RoCE, and related networking

Fluency in English

Tools

CUDA GDB

NSight Systems

cuDNN

cuBLAS

Job description

Janestreet is seeking an engineer with a focus on low-level systems programming and optimisation to join our ML team in Greater London. The role involves enhancing the performance of machine learning models, ensuring efficient large-scale training and real-time inference.

Candidates should have strong knowledge in GPU technologies and an understanding of modern machine learning techniques. Debugging, optimisation skills, and fluency in English are essential for this position.