Research Engineer, Reinforcement Learning Environments and Infrastructure

Job overview

The Research Engineer, Reinforcement Learning Environments and Infrastructure role is onsite in San Francisco, CA. Compensation is not specified. The role supports development of agent evaluation frameworks and post-training data pipelines. This engineer builds safe, verifiable environments and scalable reinforcement-learning feedback and evaluation systems for AI agents. The work helps assess agent behavior in realistic enterprise settings while improving the efficiency and reliability of distributed training and sandboxed execution.

What you'll do

• Design stateful, deterministically verifiable sandboxes for web, operating system, API, and database tasks
• implement high-throughput post-training infrastructure including RLHF, Direct Preference Optimization, and Process-Supervised Reward Models
• develop enterprise evaluation metrics and automated grading harnesses
• optimize latency and compute efficiency across distributed GPUs and sandboxed runtimes.

What we're looking for

Skills & competencies

research engineer

reinforcement learning

ai research

agent evaluation

sandbox environments

post-training pipelines

rlhf

distributed gpus

san francisco ca

onsite

python

rust

c++

go

Work arrangement

• Weekend coverage required

Why this role

The role combines systems engineering and AI research, with production AI experience and a deep systems background across languages such as Python, Rust, C++, Go, TypeScript, and CUDA sought.

About the employer

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