ML Engineer (Internship and Full-time)

What you'll do

• Optimize inference and training throughput for novel model architectures.
• Build and maintain high-performance distributed training infrastructure.
• Collaborate with researchers to turn insights into measurable improvements in model performance and understanding.

What they ask for

• Experience in deep learning or related research areas.
• Demonstrated exceptional capability in ML infrastructure work.
• Deep familiarity with PyTorch or JAX; basic familiarity with Triton, Tilelang, or similar tools.
• Ability to design and orchestrate end-to-end ML pipelines.

Nice to have

• Strong open-source contributions, thoughtful technical blog posts or work logs, or experience with large-scale pre/post-training infrastructure.
• Clear and effective verbal and written communication; ability to learn quickly.

Tools

PyTorch, JAX, Triton, Tilelang

AI focus

Supports research in mechanistic interpretability, new architectures, and pretraining science, including training, evaluating, and understanding large models.