Machine Learning Performance Engineer

Optiver is a seeking a Machine Learning Performance Engineer to join our team, focusing on a pivotal AI initiative. This role would offer the opportunity to have significant impact across Machine Learning infrastructure, training, and inference challenges to advance our futures trading strategies.

What you'll do

As a Machine Learning Performance Engineer, your key responsibilities include:

• Building scalable and robust training and inference pipelines for deep learning
• Diving into internals of open-source deep learning frameworks and enhance their functionality
• Identifying and eliminate performance bottlenecks
• Collaborating closely with researchers and other engineers
• Developing an in-depth understanding of trading systems

What you’ll get

You’ll join a culture of collaboration and excellence, surrounded by curious thinkers and creative problem-solvers. Motivated by a passion for continuous improvement, you’ll thrive in a supportive, high-performing environment alongside talented colleagues, collectively tackling some of the toughest challenges in the financial markets.

In addition, you’ll receive:

• The opportunity to work alongside best-in-class professionals from over 40 different countries
• A highly competitive compensation package
• Global profit-sharing pool and performance-based bonus structure
• 401(k) match up to 50%
• Comprehensive health, mental, dental, vision, disability, and life coverage
• 25 paid vacation days alongside market holidays
• Extensive office perks, including breakfast, lunch and snacks, regular social events, clubs, sporting leagues and more

Who you are

• Strong knowledge of low-level GPU programming with CUDA, including Tensor Cores, cooperative groups, graphs, and warp-level intrinsics
• Expertise in internals of deep-learning frameworks like PyTorch, JAX, TensorFlow, etc.
• Deep understanding of computer architecture
• Experience in C++ and Python

Nice to have

• Experience with JAX ecosystem (XLA, Flax, etc.)
• Familiarity with GPU libraries and tools such as Triton, CUB, cuDNN, and cuBLAS
• Linux system programming experience
• Experience with large-scale distributed training
• Contributions to open-source projects related to data science and machine learning