Schedule changes

Newest first. The schedule is synced from ai.engineer about hourly; only syncs that actually changed something appear below. Last updated Wed, Jul 22, 12:01 AM PDT.

Wed, Jul 22, 12:01 AM PDT
Schedule version 4945 · 561 sessions
1 changed
Rescheduled & retitled · 1
Beyond RAG: See a relational context engine reduce token burn
Speakers: +Peter Werry Brandon Waselnuk
Fri, Jul 10, 1:35 PM PDT
Schedule version 4943 · 561 sessions
1 changed
Rescheduled & retitled · 1
Beyond RAG: See a relational context engine reduce token burn
Speakers: +Brandon Waselnuk Peter Werry
Fri, Jul 10, 2:14 AM PDT
Schedule version 4945 · 561 sessions
1 changed
Rescheduled & retitled · 1
Beyond RAG: See a relational context engine reduce token burn
Speakers: +Peter Werry Brandon Waselnuk
Fri, Jul 3, 5:52 PM PDT
Schedule version 4943 · 561 sessions
+1 added
Added · 1
AMA with Theo Browne (@t3.gg)
Day 4 — Session Day 3 · 2:00pm-2:30pm · OpenAI Booth
Thu, Jul 2, 9:41 AM PDT
Schedule version 4943 · 560 sessions
2 changed
Rescheduled & retitled · 2
How Anthropic Builds: Lessons from Labs
Speakers: +swyx
Taking Reinforcement Learning Cross Datacenter
Speakers: +Nan Jiang Adam Azzam
Thu, Jul 2, 7:52 AM PDT
Schedule version 4936 · 560 sessions
+4 added2 changed
Added · 4
Battlefield setup
Day 4 — Session Day 3 · 8:30am-10:30am · Networking Room
$100,000 AIE Startup Battlefield — presented by Hyperagent
Day 4 — Session Day 3 · 10:45am-1:00pm · Networking Room
Improving AIE: Q&A and Feedback Session
Day 4 — Session Day 3 · 2:25pm-2:45pm · Expo Stage 2 NW
Improving AIE: Q&A and Feedback Session (repeat)
Day 4 — Session Day 3 · 3:20pm-3:40pm · Expo Stage 2 NW
Rescheduled & retitled · 2
Startup Battlefield
Speakers: +Joshua Xu +swyx
Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer
Title: Why Graphs? → Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer
Thu, Jul 2, 12:19 AM PDT
Schedule version 4927 · 556 sessions
+1 added6 changed−1 removed
Added · 1
Large clusters for small models
Day 4 — Session Day 3 · 1:55pm-2:15pm · Track 9
Rescheduled & retitled · 6
State of the Union: Why Local, Why Now
Speakers: +Matthew Berman
State of the Union: Why Local, Why Now
Speakers: +Matthew Berman
Local Models: Trust, Control, Optimization
Speakers: Lou
Local Models: Trust, Control, Optimization
Speakers: Lou
Compression at the Edge
Speakers: +Parth Sareen Michael Chiang
Compression at the Edge
Speakers: +Parth Sareen Michael Chiang
Removed · 1
Rob Wachen — transformer-only ASICs for inference
Day 4 — Session Day 3 · 1:55pm-2:15pm · Track 9
Wed, Jul 1, 2:37 PM PDT
Schedule version 4911 · 556 sessions
3 changed
Rescheduled & retitled · 3
The Next Run Should Be Better
Day: Day 3 — Session Day 2 → Day 4 — Session Day 3
Room: Expo Stage 3 SW → Expo Stage 4 SE
The Infinite Context Window Is a Myth: Context Engineering for AI Agents
Speakers: +Elizabeth Fuentes Leone Morgan Willis
Knowledge Systems: The New GTM Stack
Title: Lessons From Building The World's Largest Knowledge Graph → Knowledge Systems: The New GTM Stack
Wed, Jul 1, 9:53 AM PDT
Schedule version 4905 · 556 sessions
+1 added2 changed−5 removed
Added · 1
The 2026 State of AI Engineering
Day 4 — Session Day 3 · 9:00am-9:20am · Main Stage
Rescheduled & retitled · 2
Building the simulation infrastructure for practical world model use (Part 2)
Title: Building the simulation infrastructure for practical world model use → Building the simulation infrastructure for practical world model use (Part 2)
Unitree: Building Mass Produced Humanoids
Title: Unitree → Unitree: Building Mass Produced Humanoids
Removed · 5
Building Great Agent Skills: The Missing Manual
Day 4 — Session Day 3 · 9:00am-9:20am · Main Stage
From Agent Prototype to Production: Evaluation-Led Development in Microsoft Foundry
Day 4 — Session Day 3 · 1:55pm-2:15pm · Expo Stage 3 SW
Tokenomics: From AI Spend to AI Value (Working Lunch)
Day 3 — Session Day 2 · 12:30pm-1:30pm · Leadership Lounge
2026 AI Engineering Survey
Day 3 — Session Day 2 · 9:00am-9:05am · Main Stage
Cosmos / Omniverse for Robotics & World Models
Day 3 — Session Day 2 · 12:05pm-12:25pm · Track 2
Tue, Jun 30, 8:19 PM PDT
Schedule version 4895 · 560 sessions
+1 added2 changed−7 removed
Added · 1
Frontier models for the hard parts, open weights for the rest
Day 2 — Session Day 1 · 11:10am-11:30am · Expo Stage 3 SW
Rescheduled & retitled · 2
Field Guide to Fable
Title: Seeing like an Agent → Field Guide to Fable
Bringing Continual Learning into Enterprises
Title: Distillation and Continual Learning for Enterprise AI → Bringing Continual Learning into Enterprises
Removed · 7
Kimchi on the Expo Session
Day 2 — Session Day 1 · 11:10am-11:30am · Expo Stage 3 SW
Baseten Expo Session
Day 2 — Session Day 1 · 2:50pm-3:10pm · Expo Stage 3 SW
Circle Expo Session
Day 3 — Session Day 2 · 10:45am-11:05am · Expo Stage 1 NE
Daytona Expo 1
Day 3 — Session Day 2 · 1:30pm-1:50pm · Expo Stage 4 SE
Dash0 Add-On Expo Session
Day 4 — Session Day 3 · 1:55pm-2:15pm · Expo Stage 1 NE
Deepmind Expo Session 2
Day 4 — Session Day 3 · 3:20pm-3:40pm · Expo Stage 2 NW
[Braintrust Expo Session 3]
Day 4 — Session Day 3 · 2:25pm-2:45pm · Expo Stage 2 NW
Tue, Jun 30, 4:34 PM PDT
Schedule version 4886 · 566 sessions
6 changed−2 removed
Rescheduled & retitled · 6
Engineering Agency out of the Happy Path
Title: Superhuman performance is a shape, not just nines. → Engineering Agency out of the Happy Path
How Web Data Infrastructure Powers the Next Generation of AI
Time: 2:50pm-3:10pm → 3:20pm-3:40pm
Will AI predict people like we predict the weather? (alternate title “A field guide to synthetic personas for market research”)
Time: 3:20pm-3:40pm → 2:50pm-3:10pm
Learning on the job: the future of post-training
Title: Bring your own harness: flexible post-training → Learning on the job: the future of post-training
Your Stack Has a Latency Problem You Can’t See
Day: Day 2 — Session Day 1 → Day 3 — Session Day 2
Time: 11:40am-12:00pm → 2:25pm-2:45pm
Room: Expo Stage 1 NE → Expo Stage 4 SE
Why Agents Should Have Their Own Sandbox
Room: Expo Stage 1 NE → Expo Stage 3 SW
Removed · 2
Deepmind Expo Session 3
Day 3 — Session Day 2 · 1:30pm-1:50pm · Expo Stage 3 SW
Lightrun Expo Session
Day 3 — Session Day 2 · 2:25pm-2:45pm · Expo Stage 4 SE
Tue, Jun 30, 3:33 PM PDT
Schedule version 4878 · 568 sessions
+1 added2 changed−1 removed
Added · 1
ARIA, how we built autoresearch with autoresearch
Day 4 — Session Day 3 · 11:10am-11:30am · Expo Stage 2 NW
Rescheduled & retitled · 2
KV Cache-Aware Routing and P/D Disaggregation on Kubernetes: The Parts Public Benchmarks Don't Show
Speakers: Michey Mehta
Can Your Agent Hear You Now?
Time: 1:30pm-1:50pm → 3:20pm-3:40pm
Room: Expo Stage 2 NW → Expo Stage 1 NE
Removed · 1
Show your workflow: Reimagining AI PDLC with Skills
Day 2 — Session Day 1 · 11:00am-12:00pm · Leadership Lounge
Tue, Jun 30, 11:47 AM PDT
Schedule version 4870 · 568 sessions
8 changed
Rescheduled & retitled · 8
Ask YouTube — Open Q&A
Title: Don't Summarize. Sample. — How YouTube Re-Built Search for the LLM Era → Ask YouTube — Open Q&A
Video Discovery for Agentic World-Model Training
Speakers: +Rafael Levi
Recursive Model Improvement
Title: The Future of Cursor → Recursive Model Improvement
The Rise of CaaS: Context-as-a-Service for Agentic AI
Speakers: +Omer Primor TBD — Bright Data speaker
Video Discovery for Agentic World-Model Training
Speakers: +Rafael Levi
The Human Is an Async API
Room: Expo Stage 4 SE → Expo Stage 3 SW
How we Solved Agent Building
Title: We Solved Agent Building - The Evolution of Building A Successful Data Science Agent → How we Solved Agent Building
[Braintrust Expo Session 3]
Room: Expo Stage 3 SW → Expo Stage 2 NW
Tue, Jun 30, 8:01 AM PDT
Schedule version 4859 · 568 sessions
+2 added7 changed−2 removed
Added · 2
Cosmos / Omniverse for Robotics & World Models
Day 3 — Session Day 2 · 12:05pm-12:25pm · Track 2
Latent Space Live: the Inference Inflection from First Principles
Day 4 — Session Day 3 · 12:30pm-1:30pm · Expo Stage 2 NW
Rescheduled & retitled · 7
The Golden Age of AI Engineering
Title: Opening Keynote: Topic TBD → The Golden Age of AI Engineering
Will AI predict people like we predict the weather? (alternate title “A field guide to synthetic personas for market research”)
Day: Day 2 — Session Day 1 → Day 3 — Session Day 2
Time: 3:45pm-4:05pm → 3:20pm-3:40pm
Unitree
Speakers: +XiangMing Sun TBD — Unitree
The Highest Loop
Title: Three Years of AI Engineering → The Highest Loop
Building agents is trivial now, context is the next frontier
Day: Day 4 — Session Day 3 → Day 2 — Session Day 1
Time: 11:40am-12:00pm → 2:25pm-2:45pm
Room: Expo Stage 4 SE → Expo Stage 3 SW
We're the bottleneck, but we don't have to be
Title: AgentCraft: Gamifying AI Orchestration → We're the bottleneck, but we don't have to be
Can Your Agent Hear You Now?
Time: 2:25pm-2:45pm → 1:30pm-1:50pm
Room: Expo Stage 3 SW → Expo Stage 2 NW
Removed · 2
Deepmind Expo Session 1
Day 2 — Session Day 1 · 1:30pm-1:50pm · Expo Stage 2 NW
TBA
Day 3 — Session Day 2 · 12:05pm-12:25pm · Track 2
Tue, Jun 30, 12:32 AM PDT
Schedule version 4844 · 568 sessions
+11 added13 changed−3 removed
Added · 11
From approval loops to autonomous agents with Docker pt1
Day 2 — Session Day 1 · 1:30pm-1:50pm · Track 7
From approval loops to autonomous agents with Docker pt2
Day 2 — Session Day 1 · 1:55pm-2:15pm · Track 7
AgentCraft: Gamifying AI Orchestration
Day 2 — Session Day 1 · 2:25pm-2:45pm · Main Stage
From approval loops to autonomous agents with Docker pt3
Day 2 — Session Day 1 · 2:25pm-2:45pm · Track 7
Can Your Agent Hear You Now?
Day 2 — Session Day 1 · 2:25pm-2:45pm · Expo Stage 3 SW
From approval loops to autonomous agents with Docker pt4
Day 2 — Session Day 1 · 2:50pm-3:10pm · Track 7
Forward Deployed Engineering 101
Day 2 — Session Day 1 · 2:50pm-3:10pm · Track 8
From approval loops to autonomous agents with Docker pt5
Day 2 — Session Day 1 · 3:20pm-3:40pm · Track 7
Why Agents Should Have Their Own Sandbox
Day 3 — Session Day 2 · 1:30pm-1:50pm · Expo Stage 1 NE
Infra behind Krea 2 - How to train and serve at scale
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 1
An Interaction Is All You Need
Day 4 — Session Day 3 · 3:20pm-3:40pm · Expo Stage 3 SW
Rescheduled & retitled · 13
Thom Wolf keynote
Time: 9:45am-10:05am → 10:05am-10:25am
Speakers: +Olive Song
GLM-5.2: Frontier Intelligence, Open Weights.
Time: 10:05am-10:25am → 9:45am-10:05am
Rethinking Environments for Long Horizon Work
Title: Theta Software → Rethinking Environments for Long Horizon Work
How We Built the Airbyte Agent MCP Server and CLI
Speakers: +Pedro Lopez Cam Kennedy
How do you diffuse AI into the real world?
Title: Vertical Superintelligence: Making AI Work in America's Messiest Industries → How do you diffuse AI into the real world?
Speakers: Rasmus Wissmann
Reinforcement Learning without Verifiable Rewards
Title: PRIME-RL: Async & Decentralized RL Training at Scale → Reinforcement Learning without Verifiable Rewards
From Manual Drones to Autonomous Multi-Agent Missions
Speakers: +Suchet Bargoti Juraj Kabzan
Will AI predict people like we predict the weather? (alternate title “A field guide to synthetic personas for market research”)
Time: 2:50pm-3:10pm → 3:45pm-4:05pm
Preferences > Benchmarks: Model Routing for How Teams Actually Build
Title: Preferences > Benchmarks: Model Routing for How Teams Actually Build: Model Routing for How Teams Actually Build → Preferences > Benchmarks: Model Routing for How Teams Actually Build
YOLO Mode, Safely: microVM Sandboxes for Any Agent
Speakers: +Rowan Christmas Eric Jia
The Prime Intellect Stack
Title: Prime Intellect Forge → The Prime Intellect Stack
Tokens In, Engagement Out: Training LLM-Recommenders
Title: The state of LLM Recsys → Tokens In, Engagement Out: Training LLM-Recommenders
Building GTM AI Agents: Lessons from Deploying to 6,000 Users
Title: Lessons learned from building and using the internal AI platform utilized by 6000 sellers → Building GTM AI Agents: Lessons from Deploying to 6,000 Users
Removed · 3
Replicated Expo Session
Day 4 — Session Day 3 · 3:20pm-3:40pm · Expo Stage 3 SW
River: Shopify’s Slack-Native AI Agent
Day 2 — Session Day 1 · 2:25pm-2:45pm · Main Stage
Serving Diffusion Models at Scale
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 1
Mon, Jun 29, 1:42 PM PDT
Schedule version 4652 · 560 sessions
+4 added9 changed−4 removed
Added · 4
Agentic Search for Coding Agents
Day 2 — Session Day 1 · 10:45am-11:05am · Expo Stage 4 SE
AI Applications in a flash! No Dev Ops. Just code.
Day 3 — Session Day 2 · 3:20pm-3:40pm · Expo Stage 2 NW
Lessons learned from building and using the internal AI platform utilized by 6000 sellers
Day 4 — Session Day 3 · 3:20pm-3:40pm · Track 6
Dashboards are Dead
Day 4 — Session Day 3 · 3:45pm-4:05pm · Expo Stage 2 NW
Rescheduled & retitled · 9
Agentic SDLC at Uber: Building Blocks for Uber's Software Factory
Title: The AI-Native Software Factory: How Uber Agentified Every Stage of the SDLC → Agentic SDLC at Uber: Building Blocks for Uber's Software Factory
From coding to Knowledge work agents
Title: Tool Execution layer for agents → From coding to Knowledge work agents
From VLM/VLA's to Embodied Agents
Title: Perceptron Mk1 — Perceptron Inc → From VLM/VLA's to Embodied Agents
From raw documents to AI-ready data
Title: From raw documents to retrievable knowledge → From raw documents to AI-ready data
WTF Is the Context Layer? The Missing Infrastructure for Production Agents
Title: Cut Through the Context Hype: 4 Layers Your Agent Is Missing → WTF Is the Context Layer? The Missing Infrastructure for Production Agents
Simulation-Maxxing: How Nubank ships agents 20× faster with simulations
Title: Simulationmaxxing for 20x Faster Agent Dev → Simulation-Maxxing: How Nubank ships agents 20× faster with simulations
Speakers: +Aman Gupta
How Anthropic Builds: Lessons from Labs
Title: How Tag changed Labs → How Anthropic Builds: Lessons from Labs
Letting the Interns Loose — How We Accelerated AI Adoption.
Title: Building a Unified Interface For LLMs → Letting the Interns Loose — How We Accelerated AI Adoption.
Training Krea 2 - What matters in generative model training.
Title: Training Krea 2 → Training Krea 2 - What matters in generative model training.
Removed · 4
Exa Expo Session
Day 2 — Session Day 1 · 10:45am-11:05am · Expo Stage 4 SE
Runpod Expo Session
Day 3 — Session Day 2 · 3:20pm-3:40pm · Expo Stage 2 NW
Fin.ai in GTM
Day 4 — Session Day 3 · 3:20pm-3:40pm · Track 6
Composio Expo Session
Day 4 — Session Day 3 · 3:45pm-4:05pm · Expo Stage 2 NW
Mon, Jun 29, 3:43 AM PDT
Schedule version 4614 · 560 sessions
+1 added6 changed−2 removed
Added · 1
what we learned by analyzing 1M AI generated PRs
Day 2 — Session Day 1 · 1:55pm-2:15pm · Expo Stage 3 SW
Rescheduled & retitled · 6
The Data Context Layer: Why Data Engineering Agents Need More Than Code and Databases
Speakers: +Brandon Callender
Training Frontier Models to Out-Think Hackers
Title: Arithmetic (fka Operating Intelligence) → Training Frontier Models to Out-Think Hackers
Speakers: +Thom Wolf
Building the Document Context Layer for AI Agents
Title: Building LiteParse: the world's fastest PDF -> markdown parser → Building the Document Context Layer for AI Agents
Get Out of the Model's Way
Title: Defying Antigravity → Get Out of the Model's Way
Loop Engineering from first principles
Title: The Software Factory Survivor's Guide → Loop Engineering from first principles
One Designer + Al. Hundreds of Deliverables.
Title: Let My People Code → One Designer + Al. Hundreds of Deliverables.
Removed · 2
Greptile Expo Session
Day 2 — Session Day 1 · 1:55pm-2:15pm · Expo Stage 3 SW
Buildkite Expo 1
Day 2 — Session Day 1 · 2:25pm-2:45pm · Expo Stage 3 SW
Sun, Jun 28, 4:30 PM PDT
Schedule version 4602 · 561 sessions
+1 added11 changed−3 removed
Added · 1
How Reducto parsed the Epstein Files for the Viral JMail Project: The Secret Complexities of Document
Day 1 — Workshop Day · 1:15pm-2:15pm · Track 7
Rescheduled & retitled · 11
AI Security Engineer Foundations + Certificate
Title: AI Security Engineering 101 → AI Security Engineer Foundations + Certificate
AI Security Engineer Foundations + Certificate
Title: AI Security Engineering 201 → AI Security Engineer Foundations + Certificate
Kimchi on the Expo Session
Title: Kimchi Expo Session → Kimchi on the Expo Session
From Ingestion to Agents: How Leading AI Teams Build on Document Intelligence
Speakers: +Adit Abraham TBD — Reducto
Act, Confirm, or Stop? Smarter behavior for AI assistants, wearables & robots
Title: Reduce the OUCH: Smarter behavior for AI assistants and robots → Act, Confirm, or Stop? Smarter behavior for AI assistants, wearables & robots
Data and Environment Curation for Post-training LLMs
Title: Tools for data curation and post-training LLMs → Data and Environment Curation for Post-training LLMs
Codex, Behind the Harness
Day: Day 3 — Session Day 2 → Day 4 — Session Day 3
Time: 3:45pm-4:05pm → 1:30pm-1:50pm
Room: Expo Stage 2 NW → Track 8
Guide, Verify, Solve: The Engineering Discipline Agentic Development Demands
Speakers: Manish Kapur
Stop Model Shopping: Why Ownership Beats Choice in the Agent Stack
Speakers: +Pranay Bhatia George Hu
Building on the Codex Harness
Day: Day 4 — Session Day 3 → Day 3 — Session Day 2
Time: 1:30pm-1:50pm → 3:45pm-4:05pm
Room: Track 8 → Expo Stage 2 NW
Skills are new features: Building Skill-Centric Harness for Agentic Products
Title: It's a Skill Issue : Best practices in building skills that work → Skills are new features: Building Skill-Centric Harness for Agentic Products
Removed · 3
Build a Document Triage Agent with Reducto: Classify, Extract, and More
Day 1 — Workshop Day · 1:15pm-2:15pm · Track 7
AI-transforming 18K engineers, 40K repos, and an agent swarm: what worked, what didn't
Day 2 — Session Day 1 · 3:45pm-4:05pm · Leadership 1
Keycard AI Expo 1
Day 4 — Session Day 3 · 11:10am-11:30am · Expo Stage 2 NW
Sun, Jun 28, 12:35 PM PDT
Schedule version 4567 · 563 sessions
+5 added10 changed1 revived−3 removed
Added · 5
The state of LLM Recsys
Day 2 — Session Day 1 · 10:45am-11:05am · Track 7
Spotify LLM Recsys
Day 2 — Session Day 1 · 11:10am-11:30am · Track 7
LLM Recsys at DoorDash
Day 2 — Session Day 1 · 11:40am-12:00pm · Track 7
Open Q&A: LLM Recsys
Day 2 — Session Day 1 · 12:05pm-12:25pm · Track 7
Trust, But Verify: Human-in-the-Loop for Agents That Actually Matter
Day 4 — Session Day 3 · 1:30pm-1:50pm · Expo Stage 1 NE
Rescheduled & retitled · 10
Building self-learning loops for your agent
Title: How to use Arize’s newest features → Building self-learning loops for your agent
Let your agent cook: using skills to evaluate and improve your app
Title: The Cheat Sheet: get evals up and running in minute with Agent Skills → Let your agent cook: using skills to evaluate and improve your app
Data Quality is the Compute Multiplier
Title: Data curation for model training → Data Quality is the Compute Multiplier
From Signal to PR: Anatomy of a Self-Improving Agent
Title: From Alert to PR: Building an Autonomous Agent Debugger → From Signal to PR: Anatomy of a Self-Improving Agent
Tell the Robot What You Want
Day: Day 3 — Session Day 2 → Day 4 — Session Day 3
Time: 12:05pm-12:25pm → 3:45pm-4:05pm
Room: Track 2 → Leadership 1
Design multi-agent systems that actually work
Day: Day 3 — Session Day 2 → Day 4 — Session Day 3
Time: 1:30pm-1:50pm → 12:05pm-12:25pm
Will AI predict people like we predict the weather? (alternate title “A field guide to synthetic personas for market research”)
Time: 10:45am-11:05am → 2:50pm-3:10pm
Designing Evals That Earn User Trust
Day: Day 4 — Session Day 3 → Day 2 — Session Day 1
Time: 2:50pm-3:10pm → 1:30pm-1:50pm
Room: Expo Stage 2 NW → Expo Stage 3 SW
Closing Keynote
Title: Closing Keynote — TBD → Closing Keynote
Speakers: +Wei-Lin Chiang Anastasios N. Angelopoulos
Everything Is a Rollout
Title: The Future of Eval Environments → Everything Is a Rollout
Speakers: +Ryan Marten
Revived · 1
TBA
Day 3 — Session Day 2 · 12:05pm-12:25pm · Track 2
Removed · 3
TBA
Day 2 — Session Day 1 · 1:30pm-1:50pm · Track 7
TBA
Day 2 — Session Day 1 · 1:55pm-2:15pm · Track 7
TBA
Day 2 — Session Day 1 · 2:25pm-2:45pm · Track 7
Sun, Jun 28, 3:50 AM PDT
Schedule version 4524 · 560 sessions
1 changed
Rescheduled & retitled · 1
Security Track intro
Speakers: +Manoj Nair swyx
Sat, Jun 27, 6:12 PM PDT
Schedule version 4523 · 560 sessions
+2 added4 changed
Added · 2
The Future of Eval Environments
Day 3 — Session Day 2 · 3:45pm-4:05pm · Track 5
Build for the Memo, Not the Demo — Notes from 200 Investment Committees
Day 4 — Session Day 3 · 1:30pm-1:50pm · Track 3
Rescheduled & retitled · 4
Lifestyles of the AI-Native: Voice-coding, agent skills, hooks and scheduled tasks
Time: 12:10pm-1:10pm → 4:30pm-5:30pm
Room: Track 2 → Track 7
From approval loops to autonomous agents with Docker
Time: 4:30pm-5:30pm → 12:10pm-1:10pm
Room: Track 7 → Track 2
Will AI predict people like we predict the weather? (alternate title “A field guide to synthetic personas for market research”)
Day: Day 3 — Session Day 2 → Day 2 — Session Day 1
Time: 3:20pm-3:40pm → 10:45am-11:05am
Room: Track 5 → Track 7
Don't Ship Skills Without Evals
Time: 3:45pm-4:05pm → 3:20pm-3:40pm
Sat, Jun 27, 4:28 PM PDT
Schedule version 4508 · 558 sessions
+7 added6 changed−5 removed
Added · 7
Training Krea 2
Day 4 — Session Day 3 · 10:45am-11:05am · Track 1
Demo: GLM 5.2 on DGX Station — Frontier Intelligence Under Your Desk
Day 4 — Session Day 3 · 11:40am-12:00pm · Track 4
Local Models: Trust, Control, Optimization
Day 4 — Session Day 3 · 1:30pm-1:50pm · Track 4
Local Models: Trust, Control, Optimization
Day 4 — Session Day 3 · 1:55pm-2:15pm · Track 4
Compression at the Edge
Day 4 — Session Day 3 · 2:25pm-2:45pm · Track 4
Serving Diffusion Models at Scale
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 1
Compression at the Edge
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 4
Rescheduled & retitled · 6
The Next Medium: Why Real-Time Interactive Video Changes Everything for Developers
Time: 2:50pm-3:10pm → 3:20pm-3:40pm
State of the Union: Why Local, Why Now
Title: Local AI State of the Union: Part 1 → State of the Union: Why Local, Why Now
Speakers: +Joseph Nelson Michael Chiang
State of the Union: Why Local, Why Now
Title: Local AI State of the Union (cont'd) → State of the Union: Why Local, Why Now
Speakers: +Nader Khalil +Joseph Nelson +Alex Cheema +Ahmad Osman
Local AI Demos
Title: Local AI in Industry → Local AI Demos
Model Routing
Title: Local AI: Model Routing → Model Routing
Speakers: +Nader Khalil +Walden Yan +Tanay Varshney +Alex Atallah
Model Routing
Title: Local AI: Model Routing (con'td) → Model Routing
Speakers: +Nader Khalil +Walden Yan +Tanay Varshney +Alex Atallah
Removed · 5
Local AI in Industry
Day 4 — Session Day 3 · 11:40am-12:00pm · Track 4
Local AI: Best Models
Day 4 — Session Day 3 · 1:30pm-1:50pm · Track 4
Local AI: Best Models (con'td)
Day 4 — Session Day 3 · 1:55pm-2:15pm · Track 4
Local AI: Quantization & Compression
Day 4 — Session Day 3 · 2:25pm-2:45pm · Track 4
Local AI: Quantization & Compression (Con'td)
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 4
Sat, Jun 27, 11:33 AM PDT
Schedule version 4498 · 556 sessions
4 changed
Rescheduled & retitled · 4
The Dirty Secret of Forward Deployed Engineering
Title: How Forward Deployed Engineering is done at Sierra → The Dirty Secret of Forward Deployed Engineering
Don’t build agents, build environments
Title: Why Agents love Serverless Sandboxes → Don’t build agents, build environments
Closing the Loop: An Autonomous AI Research Agent
Speakers: Zubin Aysola
From Tokenmaxxing to Trusted Throughput
Title: Dealing with AI's Cost Problem without Sacrificing Innovation → From Tokenmaxxing to Trusted Throughput
Sat, Jun 27, 10:29 AM PDT
Schedule version 4477 · 556 sessions
+2 added12 changed−3 removed
Added · 2
Secure Cloud Compute
Day 2 — Session Day 1 · 3:45pm-4:05pm · Track 5
Everyone talks about document search, but what about results?
Day 4 — Session Day 3 · 1:55pm-2:15pm · Expo Stage 4 SE
Rescheduled & retitled · 12
The Autonomous Computer: Full-stack Infrastructure for Computer Use Agents
Title: Reliable Computer Use Agents require coding → The Autonomous Computer: Full-stack Infrastructure for Computer Use Agents
Gadgets: Personal app vibe coding that is actually safe
Time: 11:40am-12:00pm → 3:45pm-4:05pm
The AI-Native Software Factory: How Uber Agentified Every Stage of the SDLC
Title: Agentic SDLC at Uber → The AI-Native Software Factory: How Uber Agentified Every Stage of the SDLC
Speakers: +Adam Huda
When Will The Benchmaxxing Plague End?
Title: Beyond the Benchmark: the New Frontier of Enterprise AI Reliability → When Will The Benchmaxxing Plague End?
Everyone Gets A Software Company
Time: 3:20pm-3:40pm → 11:40am-12:00pm
Every Harness Will Become A Claw
Time: 3:45pm-4:05pm → 3:20pm-3:40pm
Build-Time vs. Run-Time: Why Your Dev Tools Will Fail in Production
Speakers: +Averi Kitsch Kurtis Van Gent
Emulated: The data for fully autonomous software engineers and companies
Title: Emulated.so → Emulated: The data for fully autonomous software engineers and companies
While You Were Generating: The Verification Gap Nobody Talked About
Day: Day 2 — Session Day 1 → Day 4 — Session Day 3
Time: 3:20pm-3:40pm → 12:05pm-12:25pm
Room: Expo Stage 1 NE → Expo Stage 2 NW
Replicated Expo Session
Time: 12:05pm-12:25pm → 3:20pm-3:40pm
Room: Expo Stage 2 NW → Expo Stage 3 SW
Bring your own harness: flexible post-training
Title: Post-Training Optimizations: Staleness, Entropy Preservation, and More → Bring your own harness: flexible post-training
Evals Track Intro
Speakers: +Aparna Dhinakaran
Removed · 3
OpenAI Cyber talk
Day 2 — Session Day 1 · 3:45pm-4:05pm · Track 5
LlamaIndex Expo Session
Day 4 — Session Day 3 · 1:55pm-2:15pm · Expo Stage 4 SE
How Reducto parsed the Epstein Files for the Viral JMail Project: The Secret Complexities of Documents
Day 4 — Session Day 3 · 2:25pm-2:45pm · Expo Stage 2 NW
Fri, Jun 26, 11:46 AM PDT
Schedule version 4443 · 557 sessions
1 changed
Rescheduled & retitled · 1
Tribal Dungeons of Global Shipping: AI Agents at Global Scale
Title: Tribal Dungeons of Global Shipping CX: AI Agents at 100K Cases a Day → Tribal Dungeons of Global Shipping: AI Agents at Global Scale
Fri, Jun 26, 4:08 AM PDT
Schedule version 4383 · 557 sessions
+15 added9 changed1 revived−12 removed
Added · 15
Building a Unified Interface For LLMs
Day 3 — Session Day 2 · 11:10am-11:30am · Track 1
Power agents with Microsoft IQ
Day 3 — Session Day 2 · 2:25pm-2:45pm · Expo Stage 1 NE
Don't Ship Skills Without Evals
Day 3 — Session Day 2 · 3:45pm-4:05pm · Track 5
Closing Keynote — TBD
Day 3 — Session Day 2 · 5:10pm-5:30pm · Main Stage
Local AI State of the Union: Part 1
Day 4 — Session Day 3 · 10:45am-11:05am · Track 4
The Agent Behind the Curtain: Building the Oz Cloud Agent Platform
Day 4 — Session Day 3 · 10:45am-11:05am · Leadership 2
Local AI State of the Union (cont'd)
Day 4 — Session Day 3 · 11:10am-11:30am · Track 4
Local AI in Industry
Day 4 — Session Day 3 · 11:40am-12:00pm · Track 4
Local AI in Industry
Day 4 — Session Day 3 · 12:05pm-12:25pm · Track 4
Local AI: Best Models
Day 4 — Session Day 3 · 1:30pm-1:50pm · Track 4
Local AI: Best Models (con'td)
Day 4 — Session Day 3 · 1:55pm-2:15pm · Track 4
Local AI: Quantization & Compression
Day 4 — Session Day 3 · 2:25pm-2:45pm · Track 4
Local AI: Quantization & Compression (Con'td)
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 4
Local AI: Model Routing
Day 4 — Session Day 3 · 3:20pm-3:40pm · Track 4
Local AI: Model Routing (con'td)
Day 4 — Session Day 3 · 3:45pm-4:05pm · Track 4
Rescheduled & retitled · 9
If we want them to do Knowledge Work, we need to design Knowledge Agents
Time: 2:25pm-2:45pm → 1:30pm-1:50pm
How to Connect AI to Billions of Legal Documents
Time: 1:30pm-1:50pm → 2:25pm-2:45pm
Where RL Will Take Search
Time: 1:55pm-2:15pm → 2:50pm-3:10pm
Your Agreements Are a Database You Can't Query. We're Fixing That
Time: 2:50pm-3:10pm → 1:55pm-2:15pm
Autoresearch for Dense Retrieval: Test-Time Compute with Frozen Embedding Models
Day: Day 2 — Session Day 1 → Day 3 — Session Day 2
Time: 3:45pm-4:05pm → 11:10am-11:30am
Room: Track 3 → Main Stage
Why Agents love Serverless Sandboxes
Title: Modal Add-On Expo Session → Why Agents love Serverless Sandboxes
Understanding is the new bottleneck
Title: On Malleable Software → Understanding is the new bottleneck
Tokens Should Have Jobs
Title: Token Should Have Jobs → Tokens Should Have Jobs
Imagination Engineering
Title: Design What People Want → Imagination Engineering
Revived · 1
TBA
Day 2 — Session Day 1 · 2:25pm-2:45pm · Track 7
Removed · 12
TBA
Day 3 — Session Day 2 · 11:10am-11:30am · Track 1
TBA
Day 4 — Session Day 3 · 10:45am-11:05am · Track 4
Rebuilding the Agent Stack for the Cloud Era
Day 4 — Session Day 3 · 10:45am-11:05am · Leadership 2
TBA
Day 4 — Session Day 3 · 11:10am-11:30am · Track 4
TBA
Day 4 — Session Day 3 · 11:40am-12:00pm · Track 4
TBA
Day 4 — Session Day 3 · 12:05pm-12:25pm · Track 4
TBA
Day 4 — Session Day 3 · 1:30pm-1:50pm · Track 4
TBA
Day 4 — Session Day 3 · 1:55pm-2:15pm · Track 4
TBA
Day 4 — Session Day 3 · 2:25pm-2:45pm · Track 4
TBA
Day 4 — Session Day 3 · 2:50pm-3:10pm · Track 4
TBA
Day 4 — Session Day 3 · 3:20pm-3:40pm · Track 4
TBA
Day 4 — Session Day 3 · 3:45pm-4:05pm · Track 4
Thu, Jun 25, 3:37 PM PDT
Schedule version 4320 · 553 sessions
+7 added118 changed−7 removed
Added · 7
Your Stack Has a Latency Problem You Can’t See
Day 2 — Session Day 1 · 11:40am-12:00pm · Expo Stage 1 NE
Warp: Building Self-Improving Agent Software Factories
Day 3 — Session Day 2 · 1:55pm-2:15pm · Expo Stage 3 SW
Vector Isn't Enough: Hybrid Search & Retrieval for AI Engineers
Day 4 — Session Day 3 · 10:45am-11:05am · Expo Stage 4 SE
Building agents is trivial now, context is the next frontier
Day 4 — Session Day 3 · 11:40am-12:00pm · Expo Stage 4 SE
Move fast and (don’t) break things
Day 4 — Session Day 3 · 12:05pm-12:25pm · Expo Stage 3 SW
Optimizing Open Models for Production Grade Inference
Day 4 — Session Day 3 · 2:25pm-2:45pm · Expo Stage 1 NE
Taking Reinforcement Learning Cross Datacenter
Day 4 — Session Day 3 · 3:45pm-4:05pm · Expo Stage 1 NE
Rescheduled & retitled · 118
Build the Right Thing: Product Engineering for Software Developers (Part 1)
Speakers: +Kent C. Dodds Kent Dodds
Build the Right Thing: Product Engineering for Software Developers — Part 2
Speakers: +Kent C. Dodds Kent Dodds
SonarQube + OpenAI: Wiring Your Team for Agentic Development
Speakers: +Killian Carlsen-Phelan
Human Connection in the Age of AI
Room: Expo Stage 2 → Expo Stage 2 NW
Getting the most out of Codex
Title: Codex Maxxing → Getting the most out of Codex
Every AI company is accidentally building a bank.
Room: Expo Stage 1 → Expo Stage 1 NE
The Enterprise Agentic Gap: When Developer-Level AI Tools Hit Millions of Lines
Room: Expo Stage 2 → Expo Stage 2 NW
How PayPal Enterprise Payments handles agent-initiated payments across ChatGPT and Google AI Mode
Room: Expo Stage 3 → Expo Stage 3 SW
Exa Expo Session
Room: Expo Stage 4 → Expo Stage 4 SE
Your LLM Stack Is a 2008 Database With Better Marketing: Why ML Security Is Dominated by Misconfiguration, Not Missing Features
Speakers: +Lovina Dmello Lovina Dmellol
Give your coding agents the power of turbogrep!
Room: Expo Stage 1 → Expo Stage 1 NE
Kimchi Expo Session
Room: Expo Stage 3 → Expo Stage 3 SW
Agents, codebases, and teams: what it actually takes to ship together
Room: Expo Stage 4 → Expo Stage 4 SE
Agentic SDLC at Uber
Speakers: Adam Huda
Agentic vs. Vector Search: An Eval-Driven Approach to Coding Agent Performance
Room: Expo Stage 2 → Expo Stage 2 NW
Agents Don't Have Coworkers, They Have Hostages
Room: Expo Stage 3 → Expo Stage 3 SW
Would your AI agent get the job? A performance review framework for enterprise agents
Room: Expo Stage 4 → Expo Stage 4 SE
Your Agent Is Lying to You About Whether It Worked
Room: Expo Stage 1 → Expo Stage 1 NE
Why building building agent quality platforms is hard.
Room: Expo Stage 2 → Expo Stage 2 NW
Can LLMs write fast multi-GPU kernels? We built a benchmark to find out.
Room: Expo Stage 3 → Expo Stage 3 SW
Self-Improving Agents That Teach the Company Back
Room: Expo Stage 4 → Expo Stage 4 SE
Every Agent, Everywhere, All at Once
Room: Expo Stage 1 → Expo Stage 1 NE
Deepmind Expo Session 1
Room: Expo Stage 2 → Expo Stage 2 NW
Stop prompting
Room: Expo Stage 4 → Expo Stage 4 SE
Who Approved That MCP Server? Governing the Tool Layer
Room: Expo Stage 1 → Expo Stage 1 NE
Voice Agents Are Mostly Invisible. Here's How to See Them.
Room: Expo Stage 2 → Expo Stage 2 NW
Greptile Expo Session
Room: Expo Stage 3 → Expo Stage 3 SW
Deploying browser agents at scale
Room: Expo Stage 4 → Expo Stage 4 SE
Beyond Golden Signals: Monitoring in the Age of GenAI
Room: Expo Stage 1 → Expo Stage 1 NE
Build agents fast with GitHub Copilot (from idea to working app)
Room: Expo Stage 2 → Expo Stage 2 NW
Buildkite Expo 1
Room: Expo Stage 3 → Expo Stage 3 SW
Continuous Engineering: Software Development for the Age of Agents
Room: Expo Stage 4 → Expo Stage 4 SE
6 Pillars of an Agentic Harness That Fixes Production Incidents
Room: Expo Stage 1 → Expo Stage 1 NE
Speakers: +Varun Krovvidi
Video Discovery for Agentic World-Model Training
Room: Expo Stage 2 → Expo Stage 2 NW
Baseten Expo Session
Room: Expo Stage 3 → Expo Stage 3 SW
Self-Driving Production: AI Wrote your Code. AI Should Fix It, Too
Room: Expo Stage 4 → Expo Stage 4 SE
From Context to Memory: Your Agents Need a Real Memory Layer
Room: Expo Stage 2 → Expo Stage 2 NW
Running a 20T-Token Data Pipeline: Infrastructure Lessons from Production
Room: Expo Stage 3 → Expo Stage 3 SW
From raw documents to retrievable knowledge
Room: Expo Stage 4 → Expo Stage 4 SE
How We Built the Airbyte Agent MCP Server and CLI
Room: Expo Stage 1 → Expo Stage 1 NE
From Chatbots to Agents: How Reducto builds for Agent Experience to Enable Real Work
Room: Expo Stage 2 → Expo Stage 2 NW
Towards Reliable Financial Agents: How a 4B Model Outsmarted a 235B Giant
Room: Expo Stage 3 → Expo Stage 3 SW
AI Enablement at Automattic: How a Remote Company Builds AI Fluency
Room: Expo Stage 4 → Expo Stage 4 SE
Harness Engineering is not Enough: Why Software Factories Fail
Speakers: +Dex Horthy Dexter Horthy
Circle Expo Session
Room: Expo Stage 1 → Expo Stage 1 NE
AI Engineering & Governance 2026 Trends
Room: Expo Stage 2 → Expo Stage 2 NW
Why AI Didn't Actually Make You Ship Faster
Room: Expo Stage 3 → Expo Stage 3 SW
Redesigning how software gets built
Room: Expo Stage 4 → Expo Stage 4 SE
Harnessing Agents: The Durable Runtime for Dynamic Workflows
Room: Expo Stage 1 → Expo Stage 1 NE
AI-Assisted Engineering: 5 Trends We're Seeing From 500+ Organizations
Room: Expo Stage 2 → Expo Stage 2 NW
The Death of Keyword Search and the Rise of Agent-Readable Catalogs
Room: Expo Stage 3 → Expo Stage 3 SW
FDE Playbook: Build an AI Support Agent and Give It a Voice
Room: Expo Stage 4 → Expo Stage 4 SE
Fault-Tolerant Training at Scale: Making Hardware Failures a Non-Event
Room: Expo Stage 1 → Expo Stage 1 NE
How to generate mergeable code with a context engine
Room: Expo Stage 2 → Expo Stage 2 NW
The Next Run Should Be Better
Room: Expo Stage 3 → Expo Stage 3 SW
AI agents don't read your policy docs. They hit your APIs.
Room: Expo Stage 4 → Expo Stage 4 SE
Your agent architecture has a half-life of 6 months
Room: Expo Stage 1 → Expo Stage 1 NE
From Stateless to Stateful: Orchestrating Real-Time Voice & Messaging Agents with Twilio and Amazon Bedrock
Room: Expo Stage 2 → Expo Stage 2 NW
Harnessing Collective Agent Intelligence for Open Science
Room: Expo Stage 3 → Expo Stage 3 SW
Prompt, Memory, Weights: The Architecture Decisions Most AI Teams Make by Accident
Room: Expo Stage 4 → Expo Stage 4 SE
Surviving Your Own Velocity: How VS Code Ships Weekly with 40 People
Room: Expo Stage 2 → Expo Stage 2 NW
Deepmind Expo Session 3
Room: Expo Stage 3 → Expo Stage 3 SW
Daytona Expo 1
Room: Expo Stage 4 → Expo Stage 4 SE
Edge-Native AI: Building Ultra-Fast Agents and MCP Servers with Spin
Room: Expo Stage 1 → Expo Stage 1 NE
Why your company needs a context graph, and how to build it
Room: Expo Stage 2 → Expo Stage 2 NW
Natively Multimodal from Step Zero
Room: Expo Stage 4 → Expo Stage 4 SE
Beyond Code Generation: API Context for Agentic Engineering
Room: Expo Stage 2 → Expo Stage 2 NW
Latency Is a Budget. Humanlike Is the Goal.
Room: Expo Stage 3 → Expo Stage 3 SW
Speakers: +Jesse Hall
Lightrun Expo Session
Room: Expo Stage 4 → Expo Stage 4 SE
Inference performance as a competitive advantage
Room: Expo Stage 1 → Expo Stage 1 NE
Building an Agent Harness for the Business, Not the Builder
Room: Expo Stage 2 → Expo Stage 2 NW
The Frontier Is Coming Home
Room: Expo Stage 3 → Expo Stage 3 SW
The Self-Improving OSS Agent Stack
Room: Expo Stage 1 → Expo Stage 1 NE
Runpod Expo Session
Room: Expo Stage 2 → Expo Stage 2 NW
Speakers: +Pranjal Jain
The Infinite Context Window Is a Myth: Context Engineering for AI Agents
Room: Expo Stage 3 → Expo Stage 3 SW
Vibe Code Safely: Introducing Gadgets
Room: Expo Stage 4 → Expo Stage 4 SE
Modular: Taming the AI Hardware Cambrian Explosion
Room: Expo Stage 1 → Expo Stage 1 NE
Codex, Behind the Harness
Room: Expo Stage 2 → Expo Stage 2 NW
Stop Renting Intelligence: The Train-to-Deploy Loop for Specialized AI
Room: Expo Stage 3 → Expo Stage 3 SW
Ray Actors, Vision Tokens, and the GIL: Engineering an SFT Data Pipeline That Keeps GPUs Busy
Room: Expo Stage 4 → Expo Stage 4 SE
AI Engineering & Governance 2026 Trends
Room: Expo Stage 1 → Expo Stage 1 NE
Your Agent Can't Tell If It's Right
Room: Expo Stage 2 → Expo Stage 2 NW
No, That's Not a Software Factory
Room: Expo Stage 3 → Expo Stage 3 SW
The Death of Developer Advocates
Speakers: Graham McBain
Beyond RAG: See a relational context engine reduce token burn
Room: Expo Stage 1 → Expo Stage 1 NE
Keycard AI Expo 1
Room: Expo Stage 2 → Expo Stage 2 NW
The Lethal Trifecta Is Already on Your Developers' Laptops
Room: Expo Stage 3 → Expo Stage 3 SW
Your AI Agent Has No Nervous System
Room: Expo Stage 4 → Expo Stage 4 SE
Speakers: +Matt Gibiec
The Art of Building Verifiers for Computer Use Agents
Room: Expo Stage 1 → Expo Stage 1 NE
Seeing the Plumbing: Profiling vLLM Speculative Decoding on NVIDIA Blackwell
Room: Expo Stage 2 → Expo Stage 2 NW
While You Were Generating: The Verification Gap Nobody Talked About
Room: Expo Stage 1 → Expo Stage 1 NE
Speakers: +Ali Adl-Tabatabai
Voice is the universal interface
Room: Expo Stage 3 → Expo Stage 3 SW
The Missing Layer in Agentic AI
Room: Expo Stage 1 → Expo Stage 1 NE
Replicated Expo Session
Room: Expo Stage 2 → Expo Stage 2 NW
Agents That Forge Their Own Tools: Self-Modifying AI in the Wild
Room: Expo Stage 4 → Expo Stage 4 SE
YOLO Mode, Safely: microVM Sandboxes for Any Agent
Room: Expo Stage 2 → Expo Stage 2 NW
Your Model is Private. Your System Isn't.
Room: Expo Stage 3 → Expo Stage 3 SW
Video Discovery for Agentic World-Model Training
Room: Expo Stage 4 → Expo Stage 4 SE
Dash0 Add-On Expo Session
Room: Expo Stage 1 → Expo Stage 1 NE
MCP doesn’t suck — your agent does
Room: Expo Stage 2 → Expo Stage 2 NW
From Agent Prototype to Production: Evaluation-Led Development in Microsoft Foundry
Room: Expo Stage 3 → Expo Stage 3 SW
LlamaIndex Expo Session
Room: Expo Stage 4 → Expo Stage 4 SE
How Reducto parsed the Epstein Files for the Viral JMail Project: The Secret Complexities of Documents
Room: Expo Stage 2 → Expo Stage 2 NW
The Human Is an Async API
Room: Expo Stage 4 → Expo Stage 4 SE
Harness Engineering: The New Core Skill for Agentic Developers
Room: Expo Stage 1 → Expo Stage 1 NE
Designing Evals That Earn User Trust
Room: Expo Stage 2 → Expo Stage 2 NW
Continuous Offensive Security the only approach in an agent-first world
Room: Expo Stage 4 → Expo Stage 4 SE
The Software Factory
Room: Expo Stage 4 → Expo Stage 4 SE
It's a Skill Issue : Best practices in building skills that work
Speakers: +Yogendra Miraje Yogi Miraje
Agent Memory Is a Solved Problem. Agent Learning Is Not.
Room: Expo Stage 1 → Expo Stage 1 NE
Deepmind Expo Session 2
Room: Expo Stage 2 → Expo Stage 2 NW
An AI Future Without the Lock-In
Room: Expo Stage 4 → Expo Stage 4 SE
Composio Expo Session
Room: Expo Stage 2 → Expo Stage 2 NW
Expo Welcome Speech
Room: Expo Stage 3 → Expo Stage 3 SW
Runway AI Film Festival
Room: Expo Stage 3 → Expo Stage 3 SW
[Braintrust Expo Session 3]
Room: Expo Stage 3 → Expo Stage 3 SW
Actionable Knowledge For Agents With Context Graphs
Room: Expo Stage 2 → Expo Stage 2 NW
Small Claws Are Beautiful: Edge Agents with NanoClaw, Raspberry Pi, and Graph Memory
Room: Expo Stage 3 → Expo Stage 3 SW
Removed · 7
Telnyx Expo Session
Day 2 — Session Day 1 · 11:40am-12:00pm · Expo Stage 1
Warp Expo Session
Day 3 — Session Day 2 · 1:55pm-2:15pm · Expo Stage 3
Elastic Expo Session
Day 4 — Session Day 3 · 10:45am-11:05am · Expo Stage 4
PLANETSCALE Expo 1
Day 4 — Session Day 3 · 12:05pm-12:25pm · Expo Stage 3
Nebius Expo Session
Day 4 — Session Day 3 · 2:25pm-2:45pm · Expo Stage 1
Modal Add-On Expo Session
Day 4 — Session Day 3 · 3:45pm-4:05pm · Expo Stage 1
[Unblocked Expo Session 3]
Day 4 — Session Day 3 · 11:40am-12:00pm · Expo Stage 4
Thu, Jun 25, 1:46 PM PDT
Schedule version 4302 · 553 sessions
+2 added6 changed
Added · 2
Your company brain will leak secrets. Here's how we stopped it for big banks and ourselves.
Day 2 — Session Day 1 · 2:50pm-3:10pm · Track 1
Why We Killed Our Multi-Agent Pipeline: Lessons From Pharma Commercial Intelligence
Day 4 — Session Day 3 · 3:45pm-4:05pm · Track 5
Rescheduled & retitled · 6
2 hr deep dive on LLM Inference at Scale — Part 1 of 2
Speakers: +Tanmay Sah
2 hr deep dive on LLM Inference at Scale — Part 2 of 2
Speakers: +Tanmay Sah
Reduce the OUCH: Smarter behavior for AI assistants and robots
Title: The Goldilocks problem: when your Robot asks too much — or acts too soon. → Reduce the OUCH: Smarter behavior for AI assistants and robots
Dealing with AI's Cost Problem without Sacrificing Innovation
Speakers: +Mingsheng Hong Sunita Verma
What If Your Chip Design Team Moved Like a Single Body?
Speakers: +Abduallah Mohamed
Distillation and Continual Learning for Enterprise AI
Speakers: +Samuel Denton Sam Denton
Thu, Jun 25, 4:04 AM PDT
Schedule version 4272 · 551 sessions
+12 added21 changed−7 removed
Added · 12
Three Years of AI Engineering
Day 2 — Session Day 1 · 9:00am-9:05am · Main Stage
Security Track intro
Day 2 — Session Day 1 · 10:25am-10:30am · Main Stage
Agentic Development Security
Day 2 — Session Day 1 · 12:05pm-12:25pm · Track 5
How Forward Deployed Engineering is done at Cognition
Day 2 — Session Day 1 · 12:05pm-12:25pm · Track 8
From Scratch to SOTA: Training a 3B State-Space Vision Model for 1.4 Billion People
Day 2 — Session Day 1 · 3:20pm-3:40pm · Track 2
2026 AI Engineering Survey
Day 3 — Session Day 2 · 9:00am-9:05am · Main Stage
Evals Track Intro
Day 3 — Session Day 2 · 10:25am-10:30am · Main Stage
From Agent Traces to Agent Simulations: The next era of agent evaluation
Day 3 — Session Day 2 · 12:05pm-12:25pm · Track 5
How long can your skills be before your agent forgets what you told it?
Day 3 — Session Day 2 · 1:30pm-1:50pm · Track 8
Training Taste
Day 3 — Session Day 2 · 1:55pm-2:15pm · Track 6
Design What People Want
Day 3 — Session Day 2 · 2:25pm-2:45pm · Track 6
Why Graphs?
Day 4 — Session Day 3 · 10:20am-10:30am · Main Stage
Rescheduled & retitled · 21
On AI and Knowledge
Time: 9:10am-9:30am → 9:05am-9:25am
Opening Keynote: Topic TBD
Time: 9:30am-9:50am → 9:25am-9:45am
Thom Wolf keynote
Time: 9:50am-10:10am → 9:45am-10:05am
GLM-5.2: Frontier Intelligence, Open Weights.
Time: 10:10am-10:30am → 10:05am-10:25am
What we learned by analyzing 1M AI-generated PRs
Time: 2:25pm-2:45pm → 12:05pm-12:25pm
Don't Summarize. Sample. — How YouTube Re-Built Search for the LLM Era
Time: 3:45pm-4:05pm → 2:25pm-2:45pm
Seeing like an Agent
Time: 9:10am-9:30am → 9:05am-9:25am
In the Land of AI Agents, the Verifiers Are King
Time: 9:30am-9:50am → 9:25am-9:45am
Perception Agents
Time: 9:50am-10:10am → 9:45am-10:05am
Research to Reality with Google DeepMind
Time: 10:10am-10:30am → 10:05am-10:25am
The Missing Layer: Design Taste in AI Agents // Stop Letting Your Agents Ship Ugly UIs
Time: 1:55pm-2:15pm → 2:50pm-3:10pm
Evals Driven-Development: Engineering a Mental Health AI Coach Ethically & Safely
Time: 2:25pm-2:45pm → 2:50pm-3:10pm
The Infinite Context Window Is a Myth: Context Engineering for AI Agents
Speakers: Clare Liguori
Building Great Agent Skills: The Missing Manual
Time: 9:10am-9:30am → 9:00am-9:20am
TCP and RDMA are Killing Inference Throughput; Homa can Fix It
Time: 9:30am-9:50am → 9:20am-9:40am
The Unreasonable Effectiveness of Separating the Task from the Model
Time: 9:50am-10:10am → 9:40am-10:00am
The Next Game Engine Won't Have a Manual
Time: 3:45pm-4:05pm → 12:05pm-12:25pm
The Signal Layer: What to Build When Anything Can Be Built
Time: 3:45pm-4:05pm → 3:20pm-3:40pm
Agent Frameworks Considered Harmful
Time: 3:45pm-4:05pm → 1:55pm-2:15pm
How Tag changed Labs
Time: 10:10am-10:30am → 10:00am-10:20am
River: Shopify’s Slack-Native AI Agent
Time: 12:05pm-12:25pm → 2:25pm-2:45pm
Removed · 7
2026 AI Engineering Survey and Arize Track Intro
Day 3 — Session Day 2 · 9:00am-9:10am · Main Stage
TBA
Day 3 — Session Day 2 · 12:05pm-12:25pm · Track 5
TBA
Day 3 — Session Day 2 · 2:50pm-3:10pm · Track 6
Tokenmaxxing
Day 3 — Session Day 2 · 5:10pm-5:30pm · Main Stage
Emil Eifrem keynote and Graphs track intro
Day 4 — Session Day 3 · 9:00am-9:10am · Main Stage
TBA
Day 4 — Session Day 3 · 10:45am-11:05am · Track 1
Beyond Prompts: Building a Multi-Agent Creative Computer That Orchestrates 5+ AI Models in Real-Time
Day 4 — Session Day 3 · 3:20pm-3:40pm · Track 1
Thu, Jun 25, 2:16 AM PDT
Schedule version 4216 · 546 sessions
+6 added14 changed−3 removed
Added · 6
River: Shopify’s Slack-Native AI Agent
Day 2 — Session Day 1 · 12:05pm-12:25pm · Main Stage
Commercial Grade-Robots for Real World Usage
Day 3 — Session Day 2 · 11:40am-12:00pm · Track 2
Building Closed-Loop Evals for a Multimodal Agent at Uber Scale
Day 3 — Session Day 2 · 11:40am-12:00pm · Track 5
Unitree
Day 3 — Session Day 2 · 1:30pm-1:50pm · Track 2
Closing Keynote
Day 3 — Session Day 2 · 4:30pm-4:50pm · Main Stage
Trends in AI
Day 3 — Session Day 2 · 4:50pm-5:10pm · Main Stage
Rescheduled & retitled · 14
RAG Needs a Map: Using GraphRAG to Retrieve Connected Context
Time: 9:00am-11:00am → 11:05am-12:05pm
AI on Your Lakehouse: Context Comes in Shapes, Not Queries
Time: 11:05am-12:05pm → 9:00am-11:00am
Where RL Will Take Search
Title: Solving Search with RL → Where RL Will Take Search
IT Admin for the AI Workforce: Why Your AI Agents Will Need Their Own IT Department
Speakers: Aman Raj
Why Off-the-Shelf AI Doesn't Understand Money
Title: AI Doesn't Understand Money (Yet) → Why Off-the-Shelf AI Doesn't Understand Money
The Building Blocks of GTM Orchestration
Title: Ramp: AI in GTM → The Building Blocks of GTM Orchestration
Rob Wachen — transformer-only ASICs for inference
Title: Gavin Uberti — transformer-only ASICs for inference → Rob Wachen — transformer-only ASICs for inference
Speakers: +Rob Wachen Gavin Uberti
Rebuilding the web for agents
Title: Nearly headless - rebuilding the web for agents → Rebuilding the web for agents
Evolution of agentic surfaces
Title: The evolution of agentic surfaces → Evolution of agentic surfaces
Show your workflow: Reimagining AI PDLC with Skills
Speakers: +Martin Harrysson +Matt Linderman +Prakhar Dixit
Tokenomics: From AI Spend to AI Value
Speakers: +Martin Harrysson +Matt Linderman +Prakhar Dixit
Tokenomics: From AI Spend to AI Value (Working Lunch)
Speakers: +Martin Harrysson +Matt Linderman +Prakhar Dixit
The Agentic Product Development Organization
Speakers: +Martin Harrysson +Matt Linderman +Prakhar Dixit
Self-Improvement of Context, Harness, and Model Weights through Reflective Optimization
Title: GEPA / reflective prompt evolution → Self-Improvement of Context, Harness, and Model Weights through Reflective Optimization
Removed · 3
What is the future of the SDLC?
Day 2 — Session Day 1 · 12:05pm-12:25pm · Main Stage
TBA
Day 3 — Session Day 2 · 4:30pm-4:50pm · Main Stage
TBA
Day 3 — Session Day 2 · 1:30pm-1:50pm · Track 2
Wed, Jun 24, 6:08 PM PDT
Schedule version 4172 · 543 sessions
+1 added20 changed2 revived−4 removed
Added · 1
GEPA / reflective prompt evolution
Day 3 — Session Day 2 · 2:25pm-2:45pm · Main Stage
Rescheduled & retitled · 20
On AI and Knowledge
Title: The Frontier Ecosystem → On AI and Knowledge
Data curation for model training
Time: 11:10am-11:30am → 10:45am-11:05am
Orchestras, not Factories
Title: Everything is Conductor → Orchestras, not Factories
IT Admin for the AI Workforce: Why Your AI Agents Will Need Their Own IT Department
Speakers: +Sarthak Aggarwal
Beyond Static Intelligence: Evaluating Continual Learning
Title: Continual Learning Bench → Beyond Static Intelligence: Evaluating Continual Learning
State of Data
Day: Day 2 — Session Day 1 → Day 3 — Session Day 2
Time: 10:45am-11:05am → 11:10am-11:30am
Computer-Use 2.0: Agents Just Got Multi-Cursor
Time: 1:55pm-2:15pm → 2:25pm-2:45pm
The Rise of CaaS: Context-as-a-Service for Agentic AI
Time: 2:25pm-2:45pm → 1:55pm-2:15pm
Speakers: +TBD — Bright Data speaker
Let My People Code
Title: Designing for AIE → Let My People Code
Building Great Agent Skills: The Missing Manual
Time: 9:30am-9:50am → 9:10am-9:30am
TCP and RDMA are Killing Inference Throughput; Homa can Fix It
Time: 9:50am-10:10am → 9:30am-9:50am
The Unreasonable Effectiveness of Separating the Task from the Model
Time: 10:10am-10:30am → 9:50am-10:10am
The Death of Developer Advocates
Time: 11:10am-11:30am → 3:45pm-4:05pm
How Juries and Librarians Can Solve GTM's AI Trust Problem
Title: The Vibe GTM Iceberg → How Juries and Librarians Can Solve GTM's AI Trust Problem
Lessons From Building The World's Largest Knowledge Graph
Speakers: +Jeffrey Wang Drew Bredvick
Reverse-Engineering the AI Buyer
Time: 3:45pm-4:05pm → 11:10am-11:30am
How Tag changed Labs
Time: 9:10am-9:30am → 10:10am-10:30am
The Messy Reality of Scale: Synthetic Data and Pre-Training at Poolside
Time: 12:05pm-12:25pm → 11:10am-11:30am
Post-Training Optimizations: Staleness, Entropy Preservation, and More
Time: 11:10am-11:30am → 12:05pm-12:25pm
Bugcrowd posttraining talk
Day: Day 3 — Session Day 2 → Day 2 — Session Day 1
Revived · 2
TBA
Day 2 — Session Day 1 · 1:55pm-2:15pm · Track 7
TBA
Day 3 — Session Day 2 · 4:30pm-4:50pm · Main Stage
Removed · 4
TBA
Day 3 — Session Day 2 · 2:25pm-2:45pm · Main Stage
Why your LLM is slow and expensive: lessons learned from running models in production
Day 4 — Session Day 3 · 1:55pm-2:15pm · Leadership 2
Scaling AI systems: where theory meets constraint
Day 4 — Session Day 3 · 2:25pm-2:45pm · Leadership 2
Quill: Event-Driven AI That Keeps Documentation Alive at Enterprise Scale
Day 4 — Session Day 3 · 3:45pm-4:05pm · Track 5
Started tracking changes
Wed, Jun 24, 11:43 AM PDT — 544 sessions on the schedule (version 4127)