Remote RL Engineer: Advanced Reinforcement Learning

Booz Allen Hamilton; United States; Hybrid; USD 99,000 - 225,000; Full time

Job summary

Booz Allen Hamilton is seeking an experienced Reinforcement Learning AI Engineer to design, implement, and train RL agents for mission-critical problems. You will work with teams across AI, data science, and software engineering to translate RL research into production-grade code and scalable training pipelines.

The role involves building agents in simulated environments using Gym or PettingZoo, applying policy optimization, value-based methods, and model-based RL, plus transitioning models to

Qualifications

• 3+ years of experience developing and training RL agents.
• Experience with Gym or PettingZoo interfaces.
• Experience with ML frameworks such as PyTorch, TensorFlow, or JAX.
• Experience developing technical solutions using Rust, Python, or C++.
• Bachelor's degree in CS, AI, or Engineering.

Responsibilities

• Design, implement, and train RL algorithms for complex decision-making problems.
• Develop scalable training pipelines using Python and modern ML frameworks.
• Build and evaluate agents in simulated environments using Gym or PettingZoo and high-fidelity simulators.
• Collaborate with domain experts to define reward structures and evaluation metrics.

Skills

RL development

ML engineering

Python

C++/Rust

Gym/PettingZoo

Education

Bachelor's in CS/AI/Engineering

Master's degree

Tools

PyTorch

TensorFlow

JAX

Gym

PettingZoo

Rust

Python

C++

Job description

Booz Allen Hamilton is seeking an experienced Reinforcement Learning AI Engineer to design, implement, and train RL agents for mission-critical problems. You will work with teams across AI, data science, and software engineering to translate RL research into production-grade code and scalable training pipelines.

The role involves building agents in simulated environments using Gym or PettingZoo, applying policy optimization, value-based methods, and model-based RL, plus transitioning models to