Markdown source

On-Device Agentic AI for the New York Times Games

Conference Context

Session Description

Traditional mobile game architectures rely on static state machines and fixed behavioral trees. Under this model, gameplay and accessibility are treated as rigid, separate systems. This results in blunt difficulty toggles, predictable character loops, and reactive features that fail to address a player's actual context. Constraint-Centric Agentic Simulation (CCAS) offers a theoretical shift. By modeling the game world as a continuous, multi-agent negotiation, accessibility and challenge become part of a single, fluid continuum. Using the JetBrains Koog framework on Android, this session explores the theory of running local agents on consumer mobile devices. We will discuss how principles of game theory, specifically dynamic negotiation and constraint satisfaction, can be used to build systems that reason over game states. Instead of executing pre-planned scripts, these agents dynamically alter their strategies. They negotiate environmental constraints to provide emergent challenges for high-skill players or organically smooth out cognitive and motor friction points for those requiring assistance. Running these theoretical models on edge hardware requires overcoming significant practical hurdles. We will break down the architecture needed to support this continuous adaptation without relying on cloud computation. We will cover how to manage memory footprints, compress state histories for rapid backtracking, and schedule local planning loops so they integrate flawlessly with the rendering engine.

Media Evidence

New York Times' Connections: A Case Study on NLP in Word Games — Shafik Quoraishee, NYT Games (speaker-match related prior/adjacent AI Engineer video; captions: English auto-captions).

These are phone-photo slide captures from the Google Photos AIE Slides album. They are supporting slide evidence and do not override official schedule fields.

Evidence Graph

This evidence graph is generated from currently linked source material: official schedule text, related video pages, cached transcripts, visible slide text, dense/reconstructed slide pages, and AI slide-classification audits.

Media Signals

Agent Reading Notes

Use these signals to refine the synopsis, topic links, people/company context, and method notes. If a source is a related external video rather than an exact official recording, keep it framed as supporting evidence.

Transcript Status

Related video transcript availability: English auto-captions. Treat this as supporting context, not a recording of this exact scheduled session unless later confirmed. Not fetched yet.

People

Supporting Slides

Slide Evidence

Synthesis

Synthesized Breakdown

On-Device Agentic AI for the New York Times Games ## Conference Context - Date/time: 2026-07-01 · 2:50pm-3:10pm - Track/room: Graphs · Track 5 - Speaker(s): Shafik Quoraishee, Joanne Song - Session type/status: sponsor · confirmed - Track: Graphs - Room: Track 5 - Session type: sponsor - Status: confirmed ## Session Description Traditional mobile game architectures rely on static state machines and fixed behavioral trees. Under this model, gameplay and accessibility are treated as rigid, separate systems. This results in blunt difficulty toggles, predictable character loops, and reactive features that fail to address a player's actual context. Constraint-Centric Agentic Simulation (CCAS) offers a theoretical shift.

Speaker And Company Context

Topics Covered

Derived Links And Source Material

Novel Concepts / Clever Methods

Evidence Boundary

This synthesis is based on the official schedule and linked source pages. It should be revisited when exact session recordings or transcript-backed secondary sources are available.