Research Engineer (ML Infrastructure)

Location: Palo Alto, United States; on-site

Company

hirify.global develops generalized physical AI and robotic systems for adaptive, reasoning-intensive work in real-world industrial environments.

What you will do

• Design and implement scalable systems for training large ML models.
• Develop distributed training systems across hundreds of GPUs, including parallelization, sharding, and compute utilization strategies.
• Improve training efficiency through attention optimization, kernel fusion, and memory management.
• Build tools for experiment tracking, monitoring, debugging, and researcher productivity.
• Debug training bottlenecks and track failures, performance, and resource utilization.
• Support model inference and cloud infrastructure for training workloads, including efficient compute resource management.

Requirements

• Strong experience building infrastructure for large-scale ML training.
• Deep understanding of training and scaling modern LLM and VLM systems.
• Proven experience setting up and scaling distributed training across hundreds of GPUs.
• Strong understanding of data, model, and pipeline parallelism.
• Strong Python programming skills.
• Expert-level proficiency in PyTorch and/or JAX, plus knowledge of attention optimization, kernel fusion, and efficient memory usage.

Nice to have

• Experience supporting inference systems in production.
• Familiarity with robotics or embodied AI workloads.
• Experience building experiment management tools for researchers.

Culture & Benefits

• Hands-on work with physical AI and robotics in industrial environments.
• Ownership-focused environment centered on solving difficult engineering problems.
• Close collaboration with modeling teams to accelerate iteration and reduce training costs.