Software Engineer - ML Infrastructure

The role

This person will lead the design and development of scalable model inference infrastructure, post-training infrastructure for reinforcement learning, and distributed ML training job execution. They will architect systems to execute complex training, inference, and evaluation workloads at scale.

What you will do

• Architect, scale, and optimize LLM post-training infrastructure for reinforcement learning and low-latency model inference services
• Design and maintain robust job execution systems for large-scale distributed ML training and fine-tuning workloads
• Build comprehensive, automated model evaluation frameworks
• Ensure high availability, low latency, and fault tolerance across GPU clusters and distributed event streams
• Collaborate cross-functionally with AI researchers, product leaders, and infrastructure teams

Required qualifications

• 5+ years of experience in software engineering, with a focused background in ML infrastructure, distributed systems, or large-scale data platform engineering
• Hands-on experience with PyTorch, model serving engines, and distributed training frameworks
• Solid background in reinforcement learning (RL) post-training infrastructure, RLHF, or automated model evaluation pipelines
• Proven ability to design and operate low-latency, high-throughput, and fault-tolerant distributed services at scale
• Strong foundation in systems design, GPU cluster utilization, and performance tuning

Benefits

• Competitive compensation aligned to the markets where team members are based
• Medical benefits
• Dental benefits
• Vision benefits
• Additional support programs

About the company

Arena is the platform for evaluating how AI models perform in the real world. Founded by researchers from UC Berkeley's SkyLab, the company builds the foundation for understanding, shaping, and benefiting from AI.