Strategic reasoning for agents

I build agents that reason strategically, and design the markets they meet in.

I am a member of the technical staff in AI Experiences.

Explore Espresso Arena Watch the keynote

{
  "protocol": "HCP/1.0",
  "entity": {
    "name": "Patrick R. Jordan",
    "type": "human",
    "role": "Member of Technical Staff",
    "team": "AI Experiences"
  },
  "interests": [
    "strategic reasoning",
    "multi-agent systems",
    "empirical game theory",
    "market design"
  ]
}

Now

When everyone has an agent, every recommendation is a move in a game.

Each agent’s best move depends on what the others do. I’m working on agents that reason about that, and on environments where useful behavior is also the winning strategy.

Strategic agents

Agents that anticipate other agents, not just predict what a user wants.

Empirical game theory

Simulate the game, solve it, learn a better response, repeat.

Market design

Access, payment, and mobility: the knobs that decide who agents work for.

Demos

Interactive · runs in your browser

Espresso Arena

Five machines. One buyer. Two offers.

Two recommender agents compete for one buyer. Watch strategies gain and lose share, run the EGTA loop epoch by epoch, and see how the incentive changes what the agents learn.

Explore the arena →