Staff Engineer: RL Post-Training & Multimodal Agents

Embedding VC; San Francisco (CA); On-site; USD 210,000 - 330,000; Full time

Job summary

Moonlake in San Francisco seeks a Member of Technical Staff to lead reinforcement learning infrastructure and model post-training. You will collaborate with the research team to improve large vision-language and code-generating agents through fine-tuning and scalable evaluation.

The role focuses on building full training systems, decisions on training needs, and preparing the company for larger-scale RL across digital and physical environments. On-site role in SF.

Qualifications

• Real-world experience training large ML models (LLM, multimodal, or code).
• Strong background in reinforcement learning and fine-tuning.
• Experience building distributed training or high-throughput systems.

Responsibilities

• Build RL and post-training pipelines for multimodal and code agents.
• Develop infra for supervised fine-tuning and reward modeling.
• Create systems to learn from agent trajectories and evaluations.
• Design rewards and verifiers for long-horizon tasks.
• Scale training and rollout across multi-GPU environments.
• Improve reliability, observability, and cost efficiency.

Skills

Reinforcement learning

Distributed training

Python

PyTorch

JAX

System design

Tools

PyTorch

JAX

Job description

Moonlake in San Francisco seeks a Member of Technical Staff to lead reinforcement learning infrastructure and model post-training. You will collaborate with the research team to improve large vision-language and code-generating agents through fine-tuning and scalable evaluation.

The role focuses on building full training systems, decisions on training needs, and preparing the company for larger-scale RL across digital and physical environments. On-site role in SF.