Remote RL Post-Training Scientist for Decentralized AI

Pluralis Research; California; Remote; USD 150.000 - 230.000; Vollzeit

Benefits dieser Stelle

Equity-heavy package

Remote-first culture

Visa sponsorship

Relocation support

Zusammenfassung

Pluralis Research is seeking an experienced engineer to own RL post-training for large language models in a decentralized setting. You will architect end-to-end pipelines from rollout ingestion to weight updates, tackling asynchronous, high-latency environments across globally distributed GPUs.

The role emphasizes production-grade Python/PyTorch engineering, research-oriented problem solving, and delivering the first decentralized post-trained models with robust evaluation.

Qualifikationen

• Hands-on RL post-training on large models (RLHF/RLVR) with systems-level experience.
• Strong Python and PyTorch engineering skills.
• Experience with asynchronous and distributed RL in high-latency environments.

Aufgaben

• Build the RL post-training stack end-to-end: rollout ingestion, reward computation, policy updates, and weight propagation.
• Invent algorithms for asynchronous, high-latency, partially trusted generation and efficient updates.
• Ship and release first decentralized post-trained models with evaluative artifacts.

Kenntnisse

Reinforcement Learning

Distributed Systems

Python

PyTorch

RL post-training

Tools

vLLM

SGLang

P2P networking

Jobbeschreibung

Pluralis Research is seeking an experienced engineer to own RL post-training for large language models in a decentralized setting. You will architect end-to-end pipelines from rollout ingestion to weight updates, tackling asynchronous, high-latency environments across globally distributed GPUs.

The role emphasizes production-grade Python/PyTorch engineering, research-oriented problem solving, and delivering the first decentralized post-trained models with robust evaluation.