Environments Intern

Reinforced Learning; United States; On-site; USD 18,000 - 28,000; Full time

Job summary

Reinforced Learning is seeking a term-long placement on the environments team. You will ship at least one environment that a customer’s agents run in, including a task, a verifier and a write-up of agent actions.

You will own an end-to-end environment with senior guidance, build verifiers, analyze rollouts, and present your findings. A CS student or self-taught candidate with Python and experience in web services or containers is preferred.

Qualifications

• Current study in computer science or a related field, or equivalent self-taught depth.
• Solid Python and some experience with web services or containers.
• Curiosity about agents, and you finish things.

Responsibilities

• Ship at least one environment that a customer’s agents run in, with a task, a verifier and a write-up of what the agents did with it.
• Own a small environment end-to-end with guidance from a senior engineer.
• Develop verifiers, write and run tests against model attempts.
• Perform rollout analysis and document failure modes.
• Present what you built and what you learned.

Skills

Python

Web services

Containers

Curiosity

Education

CS student or related field

Tools

Docker

Job description

The role

A term-long placement with the environments team. You’ll ship at least one environment that a customer’s agents run in, with a task, a verifier and a write-up of what the agents did with it.

What you’ll own

• A small environment, built end to end with guidance from a senior engineer.
• Verifiers, written and run against model attempts.
• Rollout analysis and documented failure modes.
• A presentation of what you built and what you learned.

What you bring

• Current study in computer science or a related field, or equivalent self-taught depth.
• Solid Python and some experience with web services or containers.
• Curiosity about agents, and you finish things.

Helpful, not required

• Projects involving simulators, games, test harnesses or language models.