Reinforcement Learning (RL) Engineer

About the Role

We are seeking a highly skilled Reinforcement Learning (RL) Engineer to develop, implement, and optimize RL algorithms for real-world and simulation-based applications. The ideal candidate has strong foundations in machine learning, deep learning, control systems, and hands-on experience deploying RL models in production or embedded systems.

Responsibilities

• Design, implement, and optimize RL algorithms such as PPO, SAC, TD3, DQN, A3C, TRPO, etc.
• Develop custom reward functions, policy architectures, and learning workflows.
• Conduct research on state-of-the-art RL techniques and integrate into product or research pipelines.
• Build or work with simulation environments such as PyBullet, Mujoco, IsaacGym, CARLA, Gazebo, or custom environments.
• Integrate RL agents with environment APIs, physics engines, and sensor models.
• Deploy RL models on real systems (e.g., robots, embedded hardware, autonomous platforms).
• Optimize RL policies for latency, robustness, and real-world constraints.
• Work with control engineers to integrate RL with classical controllers (PID, MPC, etc.)
• Run large-scale experiments, hyper parameter tuning, and ablation studies.
• Analyse model performance, failure cases, and implement improvements.
• Work closely with robotics, perception, simulation, and software engineering teams.
• Document algorithms, experiments, and results for internal and external stakeholders.

Skills Required

• Strong expertise in Python, with experience in ML frameworks like PyTorch or TensorFlow.
• Deep understanding of:
  • Markov Decision Processes (MDP)
  • Policy & value-based RL
  • Deep learning architectures (CNN, RNN, Transformers)
  • Control theory fundamentals
• Experience with RL libraries (stable-baselines3, RLlib, CleanRL, etc.).
• Experience with simulation tools or robotics middleware (ROS/ROS2, Gazebo).

Added Advantage

• Experience in robotics, mechatronic, or embedded systems.
• Experience with C++ for performance-critical applications.
• Knowledge of GPU acceleration, CUDA, or distributed training.
• Experience bringing RL models from simulation to real-world (Sim2Real).
• Experience with cloud platforms (AWS/GCP/Azure).

Experience

• 2–4 years of hands-on RL experience (academic or industry).
• Published RL research papers (optional but preferred).

Qualifications

• Bachelor’s/Master’s/PhD in Computer Science, Robotics, AI, Machine Learning, or related field