Agent Post-Training, Personality (AI)

TL;DR

Agent Post-Training, Personality (AI): Training and refining the behavioral and collaborative aspects of frontier AI agents with an accent on RLHF, reward modeling, and qualitative evaluation. Focus on translating subjective user experience into rigorous training signals and scalable pipelines to create thoughtful, proactive, and trustworthy agents.

Location: San Francisco, USA. Must be based in the US

Salary: $295,000 – $445,000

Company

hirify.global is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity.

What you will do

• Define and implement the collaborative "personality" of agents, focusing on judgment, proactivity, and adaptability.
• Convert qualitative behavioral observations into concrete hypotheses, evaluations, and training interventions.
• Develop and improve reward models and RL objectives to shape model behavior.
• Collaborate with human experts to create high-quality preference data and tasteful rollouts.
• Build sustainable pipelines for updating training data as behavioral standards evolve.
• Partner with product teams (ChatGPT, Codex) to integrate consumer insights into model improvements.

Requirements

• Strong technical foundations in ML, software engineering, statistics, behavioral science, or HCI.
• Experience with LLMs, post-training, RL/RLHF, reward modeling, or production ML systems.
• Ability to translate subjective product needs into falsifiable hypotheses and rigorous evals.
• Strong intuition for model behavior and user experience (taste).
• Authorized to work in the US (background checks administered per US law).

Culture & Benefits

• Opportunity to shape frontier agents used by millions of people worldwide.
• Collaborative environment working with researchers, engineers, and designers.
• Commitment to AI safety and human-centric development.
• Equal opportunity employer valuing diverse perspectives and experiences.