Note: The job is a remote job and is open to candidates in USA. 24-MAG LLC makes this full-time opportunity available through its platform; the hiring client is not identified. The role combines research and production work to build reinforcement-learning environments, training pipelines, synthetic data systems, automated evaluation frameworks, and scalable experimentation workflows.

Responsibilities

• Architect self-contained reinforcement-learning environments that capture complex real-world tasks
• Design reward functions, verifiers, evaluation logic, and supporting environment components
• Structure environments to support reliable experimentation and measurable model improvement
• Translate research objectives into technically rigorous RL workflows
• Ensure environments remain reproducible, testable, and suitable for iterative model development
• Design and scale episode pipelines and multi-component training processes
• Build reproducible experimentation workflows supporting reinforcement-learning research
• Develop systems for running, tracking, and analysing large-scale training experiments
• Improve reliability and efficiency across RL training infrastructure
• Support rapid iteration between environment design, training, evaluation, and model refinement
• Build automated data-generation systems using synthetic data to accelerate training cycles
• Develop AI-driven evaluation and quality-assurance systems for grading, validation, and feedback
• Establish automated feedback loops that improve training-data and model quality
• Design verification systems that distinguish strong model behaviour from superficially plausible outputs
• Apply rigorous quality standards throughout data-generation and evaluation pipelines
• Fine-tune and optimise open-source reinforcement-learning and machine-learning models
• Apply internally generated datasets and custom training strategies to improve model performance
• Develop benchmarking frameworks measuring capability, robustness, and data quality
• Analyse model behaviour across internal and external evaluation environments
• Contribute to the development, release, and interpretation of research evaluations and benchmark results

Skills

• Deep professional or research experience in reinforcement learning
• Strong understanding of RL environment design, reward structures, training dynamics, and evaluation
• Demonstrated experience building and scaling RL systems, training pipelines, or experimentation frameworks
• Strong experience with automation and synthetic data-generation workflows
• Familiarity with automated evaluation, model validation, and quality-assurance systems
• Experience fine-tuning and evaluating open-source machine-learning models
• Strong technical writing and communication skills
• Ability to operate effectively in fast-paced, research-driven, and highly collaborative environments
• Experience publishing benchmarks, evaluations, or research artifacts
• Familiarity with modern evaluation ecosystems and benchmarking frameworks
• Experience with scalable infrastructure supporting large-scale RL experimentation

Benefits

• Fully remote

Company Overview