Staff Research Scientist, Riot Games

• MS in Computer Science or equivalent with 6+ years relevant experience or BS in Computer Science or equivalent with 8+ years relevant experience
• 3+ years experience in deep learning
• 2+ years experience in reinforcement learning
• PyTorch or TensorFlow
• Familiarity with GCP or AWS
• Ability to read research papers and implement similar ideas

Desirable

• Tech Research performs applied research to accelerate the pace of building AI enabled applications for our game teams and to establish AI as a core capability at Riot
• In this role, you will lead the engagement with an R&D game team to develop AI agents that can play the game like a human at a high skill level
• You will be responsible for the end-to-end development of the agent as well as any infrastructure required to train the agent or integrate the agent into features required by the game teams
• Propose, implement, and evaluate algorithms that can play games like a human at a high skill level
• Work with game teams to propose, define, and develop AI enabled features (e.g. practice opponents, metagame analysis, player coaching)
• Define and gather game state from the game server needed for training
• Optimize a distributed deep learning training infrastructure including algorithms, hyperparameters, CPU, memory, and I/O
• Inform and manage stakeholders on game teams