1.2 Multi-agent AI Research Engineer: Scalable Robot Fleet Coordination

What You Will Get To Do

• Develop fundamental algorithms for multi-agent coordination (including differentiable game theory, mean-field control, and decentralized optimization) to enable fleets of autonomous robots to operate in real-world, high-stakes environments.
• Design computationally tractable formulations of multi-agent Nash equilibria, Stackelberg games, and cooperative decision-making strategies, ensuring robust and scalable decision-making across heterogeneous robotic teams.
• Build predictive models for multi-agent interaction dynamics, leveraging graph-based learning and control-theoretic formulations to drive efficient coordination in dynamic, adversarial, and uncertain settings.
• Develop distributed inference and control policies using neural PDEs, mean-field game-theoretic methods, and scalable stochastic optimization for real-time at-scale robotic interaction.
• Bridge theory with deployment—integrate multi-agent planning, auction-based task allocation, and decentralized multiagent reinforcement learning (MARL) into hardware-in-the-loop robotic systems operating at scale.
• Push the limits of explainability in multi-agent AI, ensuring tractability, convergence guarantees, and real-world feasibility while maintaining risk-aware and uncertainty-resolving decision-making.
• Collaborate across teams to transition multi-agent models from high-fidelity simulations to real-world deployments, working alongside robotics engineers, AI/ML researchers, and field roboticists to ensure seamless real-world operation.

What You Have

• Ph.D. in Applied Mathematics, Game Theory, Control Theory, Computer Science, or a related field, with expertise in multi-agent decision-making and coordination algorithms.
• Deep understanding of game-theoretic methods—including differential games, Nash equilibria, mean-field games, and Stackelberg equilibria—with a focus on scalability and tractability.
• Experience with multi-agent RL (MARL) and distributed optimization for large-scale robotic coordination in imperfect information settings.
• Hands-on experience implementing multi-agent algorithms in real-time robotic or AI-driven systems, with exposure to hardware constraints, real-world latency, and stochastic disturbances.
• Proficiency in Python, C++, or Julia, with experience in optimization libraries (e.g., CVXPY, Gurobi, JAX), reinforcement learning frameworks (e.g., RLlib, Acme), and multi-robot simulators.
• Experience working with large-scale robotic coordination (e.g., drone swarms, autonomous fleets, or industrial automation systems) is a strong plus.
• Ability to transition theoretical insights into scalable, field-deployable systems, ensuring robustness under uncertainty and adaptability to real-world constraints.