Member of Technical Staff, Post-Training (Fully Remote)

Cohere; United States; Hybrid; USD 150,000 - 210,000; Full time

Benefits offered by this job

Remote-friendly

Job summary

Cohere is seeking a senior RL researcher in the United States to design, run, and analyze large-scale RL post-training experiments for LLMs, including RLHF.

You will explore out-of-distribution generalization, reward design, and task diversity, and build held-out evaluations to catch reward hacking and benchmark overfitting. This role is remote-friendly with offices worldwide, fostering a global research team.

Qualifications

• Hands-on LLM training experience required.
• Experience applying RL to LLMs (PPO, GRPO, RLHF).
• Strong research and software engineering skills.
• Excellent communication and data quality intuition.

Responsibilities

• Design, run, and analyze large-scale RL post-training experiments.
• Research methods for out-of-distribution generalization and reward design.
• Build held-out evaluations to detect reward hacking and overfitting.

Skills

LLM training experience

Reinforcement learning

Experimentation rigor

Software engineering

Communication skills

Tools

PPO

RLHF

GRPO

Job description

• Design, run, and analyse large-scale RL post-training experiments (e.g., RLHF, RLVR).
• Research and develop methods for out-of-distribution generalization, including reward design and task diversity.
• Build held-out evaluations to catch reward hacking and benchmark overfitting.

Qualifications:

• Strong research and software engineering skills with hands-on LLM training experience.
• Experience applying RL to LLMs (PPO, GRPO, RLHF) and rigorous experimental practice.
• Excellent communication and data quality intuition.

How We Work:

• Remote-friendly with offices in Toronto, London, NYC, SF, Montreal, Paris, Berlin, and Seoul.
• No minimum in-office requirement; we care most about hiring exceptional people regardless of location.

Cohere

Cohere is a leading security-first enterprise AI company building cutting-edge foundation models and end-to-end products for businesses. We are a global team of researchers, engineers, and designers passionate about their craft, headquartered in Toronto with offices worldwide.