RL Environment Engineer: Build & Benchmark Learning Agents

mecka; San Francisco (CA); On-site; USD 140,000 - 210,000; Full time

Job summary

Mecka AI in San Francisco builds data infrastructure for robotics and embodied AI. The Research Engineer, RL Env role focuses on translating real tasks into computable environments, implementing reset behavior, observations, actions, state transitions, and measurable outcomes for agents.

You will develop rewards, run learning experiments, ensure reliable evaluation, investigate failures, and collaborate with domain experts and engineers to turn research prototypes into reusable environments and

Qualifications

• Hands-on experience formulating RL problems and building environments.
• Strong programming and debugging skills to implement research systems that others can run.
• Sound experimental design and statistical reasoning, including controlled comparisons and evaluation splits.
• Ability to reason about environment dynamics, reward design and agent behavior.
• Independent research judgment and clear communication with domain experts and engineers.

Responsibilities

• Build environments: define tasks, observations, actions and state transitions; implement reset and termination.
• Develop rewards and scoring: translate objectives into feedback and evaluation criteria.
• Run learning experiments: implement baseline agents, train and compare policies, run controlled experiments.
• Make evaluation reliable: separate training/held-out tasks, check leakage, report uncertainties and scores.
• Investigate failures: inspect trajectories and learning behavior to identify root causes.
• Build with the team: collaborate to turn research prototypes into reusable environments and tools.

Skills

RL environment

Programming & debugging

Experimental design

Analysis of environment dynamics

Communication & judgment

Job description

Mecka AI in San Francisco builds data infrastructure for robotics and embodied AI. The Research Engineer, RL Env role focuses on translating real tasks into computable environments, implementing reset behavior, observations, actions, state transitions, and measurable outcomes for agents.

You will develop rewards, run learning experiments, ensure reliable evaluation, investigate failures, and collaborate with domain experts and engineers to turn research prototypes into reusable environments and