OCT 16 – 18, 2026

AI.Engineer Sandbox — NYC

EXPLORE THE PROGRAM

Every session,
search and filter.

Search by title or speaker, then narrow by track, format, or location. All times PDT.

AI Agents · Keynote

Opening keynote: the year agents grew up

Fri, Oct 16 · 9:00 AM PDTMain Stage

Where the last twelve months actually landed, and what shipped.

Platforms · Talk

Caching at the edge without losing your mind

Fri, Oct 16 · 10:00 AM PDTMain Stage

What a CDN can and cannot do for an inference path.

AI Agents · Talk

Evals that survive contact with users

Fri, Oct 16 · 10:30 AM PDTMain Stage

Offline scores lie. Here is what we replaced them with.

Platforms · Workshop

⚠ Demo conflict A — Platform deep dive

Fri, Oct 16 · 11:00 AM PDTWorkshop A

Double-booked into Workshop A on purpose, so the Conflicts tab has something to resolve.

Security · Workshop

⚠ Demo conflict A — Vector search at scale

Fri, Oct 16 · 11:30 AM PDTWorkshop A

The other half of the room clash.

Platforms · Talk

Retrieval is not a database problem

Fri, Oct 16 · 1:00 PM PDTWorkshop B

Before: after.

AI Agents · Panel

⚠ Demo conflict B — Agent evals live

Fri, Oct 16 · 2:00 PM PDTStudio

Same speaker as the Guardrails workshop, at the same time, on purpose.

Security · Workshop

⚠ Demo conflict B — Guardrails that do not annoy anyone

Fri, Oct 16 · 2:15 PM PDTAtrium

The other half of the speaker clash.

AI Agents · Talk

Cost controls for long-running agents

Sat, Oct 17 · 9:30 AM PDTMain Stage

Budgets, caps and the alert that actually pages someone.

Platforms · Talk

Observability for prompt pipelines

Sat, Oct 17 · 10:15 AM PDTWorkshop A

Traces that survive a prompt rewrite.

Community · Talk

Scaling human review

Sat, Oct 17 · 11:00 AM PDTWorkshop B

What we learned staffing a review queue that never emptied.

Community · Panel

Closing panel: what we would build next

Sat, Oct 17 · 4:00 PM PDTMain Stage

Two speakers, one room, no conflicts.