Reinforcement Learning Inference Engineer

Job overview

The Reinforcement Learning Inference Engineer role is onsite in Palo Alto, CA. Compensation is $180,000–$440,000 per year. SpaceXSI is building a super-intelligence stack, including compute, models, and applications, to help humanity understand the universe. The engineer designs and optimizes inference infrastructure for reinforcement-learning workloads, from small-scale experiments to production training runs. The work addresses performance bottlenecks in large-scale systems and supports implementation of new reinforcement-learning techniques with the modeling team.

What you'll do

• Design and optimize the inference stack for reinforcement-learning workloads
• analyze, profile, and address performance bottlenecks in large-scale RL systems
• work with the modeling team to implement novel RL techniques and algorithms.

What we're looking for

Skills & competencies

reinforcement learning

inference engineer

llm inference

distributed systems

low-precision computing

python

c++

rust

pytorch

jax

cuda

palo alto ca

$180k–$440k/year

spacexsi

Work arrangement

• Weekend coverage required

Benefits & perks

• Equity
• medical, vision, and dental coverage
• 401(k)
• short- and long-term disability insurance
• life insurance
• discounts and other perks.

Why this role

Focus on low-precision reinforcement-learning training and inference; experience with distributed systems and LLM inference is required. Python, C++, Rust, PyTorch, JAX, and CUDA are listed technologies; quantization and inference engines such as SGLang and vLLM are preferred.

About the employer

SpaceXSI is hiring for this role. Industry: Research and Development in the Physical, Engineering, and Life Sciences (except Nanotechnology and Biotechnology). Sector: 54.