I will design deep reinforcement learning algorithms

Hello!

I am a Ph.D. student working on Deep Reinforcement Learning (RL). I have experience on a variety of academic and industrial-level projects, and have presented my work at top AI conferences.

Here, I offer help in building RL agents that can solve almost any real-world problem, including multi-objective and multi-agent, preference-based, and Human-guided RL. I am an expert in the following areas:

• Model-based or model-free RL algorithms
• Simple or complex Markov decision process (MDP) problems
• On-policy and off-policy RL algorithms, such as:
• Q-learning, SARSA, CQL, DQN, A2C, TRPO, PPO, SAC, TD3, and other model-free RL algorithms
• Any custom or simulated environments, such as traffic light controllers (SUMO), data center controllers, MuJoCo, Unity, PyBullet, and OpenAI Gym
• Multi-agent RL problems, such as Independent DQN, Independent PPO, MAPPO, QMIX, VDN, and other value-factorization methods
• Preference-based Reinforcement learning such as PbRL with synthetic and human feedback

I am comfortable working with the following languages and tools:

• Python
• TensorFlow
• Keras
• PyTorch
• NumPy
• Pandas
• Matplotlib

Please feel free to message me to discuss your work.

Best,

Umer