RL Engineer

Ethos; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Benefits offered by this job

Meaningful equity

Job summary

Ethos is building AI to make humans irreplaceable, connecting experts to paid opportunities and advisory roles. In this early-stage role, you will design RL environments and data pipelines that feed post-training workflows, and you will help build agentic infrastructure that sources, screens, and routes domain experts at scale.

You will craft evaluation frameworks to monitor data quality and model performance, contributing to a product backed by top investors.

Qualifications

• Hands-on experience designing RL environments or data for post-training pipelines (RLHF, RLAIF, preference data, evals).
• Startup experience — shipped in resource-constrained environments with ownership.
• Strong background building AI/agentic applications (LLM pipelines, tool use, evals).
• Generalist instincts — youd rather solve the problem than protect your lane.
• Comfortable across the stack: backend, infra, and ML tooling.

Responsibilities

• Design RL environments and data generation pipelines feeding post-training workflows.
• Build agentic infrastructure to source, screen, and route domain experts at scale.
• Develop evaluation frameworks to close the loop between data quality and model performance.

Skills

RL environments

Data pipelines

RLHF/RLAIF

Agentic AI apps

Startup experience

Generalist mindset

Backend/infrastructure

Job description

About This Opportunity

About Ethos

We\'re building AI to make humans irreplaceable. Ethos is the platform that matches human expertise to paid opportunities, from AI model training to advisory roles and full-time positions, backed by a16z, General Catalyst and other top investors.

What You\'ll Work On

• Designing and building RL environments and data generation pipelines that feed into post-training workflows
• Agentic infrastructure that sources, screens, and routes domain experts at scale
• Evaluation frameworks to close the loop between data quality and model performance

This is an early-stage role with meaningful equity. If you\'ve built things that matter at companies that didn\'t have all the answers, we\'d love to talk.

Key Requirements

• Hands-on experience building RL environments or generating data for post-training (RLHF, RLAIF, preference data, evals) — this is central to the role
• Startup experience — you\'ve shipped in resource-constrained environments and taken real ownership
• Strong background building AI/agentic applications (LLM pipelines, tool use, evals)
• Generalist instincts - you\'d rather solve the problem than protect your lane
• Comfortable across the stack: backend, infra, and ML tooling