AAIF Logo
Run of Show
SATURDAY, JANUARY 23, 2027 • DALLAS, TX • AAIF DALLAS IN COOPERATION WITH IEEE

Agentic AI Day

Production readiness, architectural patterns, multi-agent interoperability, and enterprise security. Moving from single-prompt wrappers to resilient, autonomous multi-agent systems.

Dual Afternoon Tracks Hands-On Coding Labs MCP & A2A Protocols Technical Co-Sponsor: IEEE
Registration Notice: Environment setup instructions, code repository access, and prerequisites will be emailed directly to registered participants prior to the event.
Audience
Staff & Architects
Engineers, AI Researchers & Leads
Morning Structure
Dual Keynotes & Core
Stateful patterns & discovery
Afternoon Tracks
Track A & Track B
Auditorium Deep-Dives + Lab
Hands-On Labs
Prerequisites Sent
Setup guides provided upon registration
Section 1

Event Architecture & Objectives

Engineered for systems architects, staff engineers, AI researchers, and technical leaders bridging the gap between toy prompts and production agent clusters.

TRACK A • MAIN AUDITORIUM 1:15 PM – 4:30 PM

Deep-Dive Systems Engineering

Focused on the hardest infrastructure challenges when scaling agentic workloads into enterprise production environments:

  • MicroVM & Wasm Sandboxes: AI-DLC runtime isolation, AWS Firecracker execution, memory safety, and sub-millisecond cold starts.
  • Gateways & Vulnerability Defense: Indirect prompt injection guards, confused deputy mitigation, egress filtering, and token firewalls.
  • OpenTelemetry Tracing & Audit Trees: Non-deterministic span propagation and cryptographically signed Evidence Artifacts.
  • Trajectory & Causality Metrics: Decision-tree efficiency, cyclic loop detection, and drift analysis vs. model hallucination.
TRACK B • HANDS-ON LAB 1:15 PM – 4:30 PM

Practical Implementation & Protocols

Facilitated, code-first workshops where engineers build, harden, and inter-operate production-grade agents in real-time:

  • Production MCP Server Construction: Exposing databases & ticket systems, Server-Sent Events (SSE), and Streamable HTTP auth.
  • Hardening MCP against Tool Poisoning: Defending against rug-pull attacks, hijacked schemas, and enforced gateway policies.
  • Agent Cards & A2A Discovery: Publishing /.well-known/agent.json and orchestrating async state lifecycles.
  • End-to-End Orchestrator Pipeline: Wiring orchestrators to remote A2A peers backed by private MCP servers using unified trace IDs.
Section 2

Full-Day Run-of-Show Schedule

Time Zone: Central Time (CT) • Strict Stage Timers Enforced
08:00 – 09:00 Main Foyer
General Operations

Doors Open & Registration

Badge pickup, coffee, Wi-Fi onboarding, and A/V soundchecks across all stages.

09:00 – 09:45 Keynote 1 • Main Auditorium
Plenary Session

The Production Shift

Moving from Single-Prompt Wrappers to Autonomous Multi-Agent Systems. Transitioning to resilient, stateful agent networks, deterministic checkpoints, and production uptime engineering.

Speaker: Keynote Speaker 1 (Announcing Soon)
09:45 – 10:30 Keynote 2 • Main Auditorium
Plenary Session

The Internet of Agents

Cross-Organizational Communication & Protocols. Open standards, decentralized ecosystems, dynamic capability registries, and cross-boundary negotiation.

Speaker: Keynote Speaker 2 (Announcing Soon)
10:30 – 10:45 Morning Break (Coffee Refresh & Speaker Mic Swaps) Foyer & Exhibit Area
10:45 – 11:30 Main Auditorium
Technical Foundation

Architecture Patterns & Workflows

Orchestration vs. choreography, state management, human-in-the-loop (HITL) checkpoints, retry storms, and durable multi-day execution engines.

Lead: Architecture & Workflows Lead
11:30 – 12:15 Main Auditorium
Security & Trust

Agent Discovery, Identity & Trust

Cryptographic agent identities, capability registries, SPIFFE/SPIRE attestation, and verifying third-party agents prior to context handoff.

Lead: Identity & Security Lead
12:15 – 01:15 Networking Lunch (Catered Buffet & Dietary Accommodations) Dining Hall
⚡ AFTERNOON PARALLEL TRACKS COMMENCE: TRACK A (MAIN AUDITORIUM) & TRACK B (HANDS-ON LAB)
01:15 – 02:00 TRACK A • AUDITORIUM

Sandboxing & Runtimes

MicroVMs, Wasm & AI-DLC Isolation. Memory safety, cold-start mitigation, and hypervisor-level isolation for untrusted code execution.

Lead: Track A Systems Lead
01:15 – 02:00 TRACK B • HANDS-ON LAB

Lab 1: Production MCP Servers

Tool Definitions & Auth. Exposing enterprise resources via standard MCP interfaces, transport security (SSE/Streamable HTTP), and scoped OAuth.

Facilitator: Track B Lab Facilitator
02:00 – 02:45 TRACK A • AUDITORIUM

Gateways & Vulnerabilities

Prompt Injection & Egress Filtering. Defense-in-depth against indirect injection, confused deputy scenarios, and real-time PII redaction.

Lead: Track A Security Lead
02:00 – 02:45 TRACK B • HANDS-ON LAB

Lab 1 Continued: Hardening MCP

Mitigating Tool Poisoning. Defending against rug-pull attacks, schema manipulation, and enforcing gateway security policies programmatically.

Facilitator: Track B Lab Facilitator
02:45 – 03:00 Afternoon Break (Refreshments & Lab 2 Connectivity Checks) Foyer & Lab Room
03:00 – 03:45 TRACK A • AUDITORIUM

Observability & Tracing

OpenTelemetry & Evidence Artifacts. Mapping non-deterministic execution paths into structured trace trees and verifiable audit logs.

Lead: Track A Systems Lead
03:00 – 03:45 TRACK B • HANDS-ON LAB

Lab 2: Multi-Agent A2A Interop

Agent Cards & Task Lifecycles. Publishing /.well-known/agent.json, handling asynchronous task state changes, and negotiating capabilities.

Facilitator: Track B A2A Facilitator
03:45 – 04:30 TRACK A • AUDITORIUM

Trajectory & Causality

Path Grading & Drift Analysis. Assessing decision-tree efficiency, detecting cyclic loops, and evaluating context-causality metrics.

Lead: Track A Systems Lead
03:45 – 04:30 TRACK B • HANDS-ON LAB

Lab 2 Continued: End-to-End Workflow

Full Pipeline Integration. Connecting an orchestrator agent to remote A2A peers backed by private MCP servers using distributed trace IDs.

Facilitator: Track B A2A Facilitator
04:30 – 05:00 Main Auditorium & Foyer
Community Wrap-Up

Closing Remarks & Happy Hour

Both tracks reconvene for final key takeaways, feedback collection, community announcements, and informal networking.

Section 3

Technical Session Scope & Speaker Guide

Speakers and session leads must align their content with rigorous production depth, actionable code, and concrete architectural trade-offs.

01

Architecture Patterns & Workflows

Scope: Deterministic state machines vs. dynamic loops. Handling retry storms, infinite loops, and persistence models for multi-day runs.

Speaker Directive: Avoid abstract conceptual diagrams. Demonstrate concrete state recovery and HITL pause/resume logic using durable execution engines.
02

Internet of Agents & Discovery

Scope: Cross-organizational agent-to-agent negotiation protocols, dynamic catalogs, zero-trust perimeters, and capability matching.

Speaker Directive: Address practical cross-domain boundaries, untrusted metadata payloads, and latency costs of remote agent discovery.
03

Identity & Trust (SPIFFE/SPIRE)

Scope: User-delegated OAuth vs. machine-to-machine identities, cryptographic attestation, SPIFFE/SPIRE, and ephemeral credentials.

Speaker Directive: Highlight failure modes where agents escalate privileges via shared or over-permissioned service account keys.
04

Sandboxing & AI-DLC Runtimes

Scope: Docker containers vs. MicroVMs (AWS Firecracker) vs. Wasm isolation for executing untrusted agent code safely.

Speaker Directive: Benchmark startup latency, memory overhead, and file system isolation during runtime tool generation.
05

Gateways & Semantic Guardrails

Scope: Centralized proxy layers, token budget management, dynamic PII masking, and semantic policy checks prior to outbound calls.

Speaker Directive: Focus on real-time latency overhead introduced by guardrail proxies and async policy evaluation patterns.
06

Agentic Vulnerability Analysis

Scope: Threat modeling indirect prompt injection in RAG, confused deputy exploits, tool-poisoning, and lateral movement.

Speaker Directive: Walk through real-world attack trees and demonstrate concrete mitigation at application and gateway layers.
07

Observability & Evidence Artifacts

Scope: Mapping OpenTelemetry (OTel) across async decision trees. Generating cryptographically signed Evidence Artifacts for compliance.

Speaker Directive: Provide concrete JSON schema structures for span context propagation across heterogeneous agent frameworks.
08

Trajectory & Causality Evaluation

Scope: Evaluating non-deterministic agent runs: path optimality, redundant tool calls, trajectory drift, and context causality checks.

Speaker Directive: Contrast classical offline evaluation pipelines with continuous production trajectory monitoring.
Section 4

Hands-On Workshop Guide (MCP & A2A)

The afternoon hands-on track provides practical, code-first workshops focused on real-world MCP and A2A implementations.

LAB PREREQUISITES & PARTICIPANT PREPARATION

Lab Prerequisites Provided Upon Registration

To ensure all participants are properly configured for the hands-on sessions, repository access, dependency setup instructions, and environment prerequisites will be sent directly to registered attendees based on their confirmed registrations prior to the event.

Detailed Setup Guide Sent Post-Registration
Technical TAs Available for On-Site Assistance
LAB 1 (01:15 – 02:45) Model Context Protocol

Production MCP Servers & Hardening

Constructing enterprise-ready MCP endpoints and defending against tool poisoning and rogue parameter injection.

Module 1: Structured Resources & Tools

Expose databases and internal ticketing endpoints via standardized MCP JSON-RPC tool declarations.

Module 2: Transport Security & OAuth

Secure transport via Server-Sent Events (SSE) and Streamable HTTP with scoped OAuth tokens.

Module 3: Defending Against Tool Poisoning

Implement defensive schema validation to guard against rug-pull attacks designed to hijack model tool selection logic.

mcp-server-config.json
{
  "name": "enterprise-tickets-mcp",
  "version": "1.0.0",
  "transport": "sse",
  "auth": {
    "type": "oauth2_bearer",
    "scopes": ["tickets:read", "tickets:write"]
  },
  "guardrails": {
    "sanitize_tool_descriptions": true,
    "max_payload_bytes": 1048576
  }
}
LAB 2 (03:00 – 04:30) Agent-to-Agent Protocols

Multi-Agent A2A Interoperability

Implementing cross-agent communication protocols, asynchronous task states, and federated peer delegation.

Module 1: Standard Agent Cards

Publish and parse standard capability contracts at /.well-known/agent.json.

Module 2: Async Task Lifecycle States

State machines for submitted, working, input-required, completed, and failed.

Module 3: Attestation & End-to-End Orchestration

Enforce cryptographic verification before wiring orchestrator agents to remote A2A peers backed by MCP servers.

/.well-known/agent.json
{
  "agent_id": "aaif.dallas.billing-agent-01",
  "name": "Enterprise Billing Agent",
  "protocol_version": "a2a/v1.0",
  "capabilities": ["invoice.generate", "ledger.reconcile"],
  "attestation": {
    "spiffe_id": "spiffe://aaif.internal/agent/billing",
    "public_key": "ecdsa-sha2-nistp256 AAAAE2V..."
  }
}
Section 5

Featured Speakers & Technical Leads

World-class systems architects, security researchers, and lead maintainers driving autonomous agent infrastructure from concept to production.

Keynote Speaker 1
KEYNOTE 1

Keynote Speaker 1

Announcing Soon

AI Systems & Architecture

The Production Shift: Single-Prompt Wrappers to Stateful Multi-Agent Clusters

Pioneering reliable, stateful agent networks, deterministic checkpoints, and multi-day long-running workflow persistence models.

09:00 – 09:45 CT Main Auditorium
Keynote Speaker 2
KEYNOTE 2

Keynote Speaker 2

Announcing Soon

Protocols & Open Standards

The Internet of Agents: Cross-Organizational A2A Negotiation

Architecting decentralized capability catalogs, zero-trust perimeter federation, and inter-organizational task delegation.

09:45 – 10:30 CT Main Auditorium
Track A Systems Lead
TRACK A LEAD

Track A Systems Lead

Announcing Soon

MicroVMs & Runtimes

Sandboxing & AI-DLC Runtimes: MicroVMs & Wasm

Analyzing hypervisor isolation, AWS Firecracker microVM cold-start mitigation, and memory safety during runtime code synthesis.

01:15 – 02:00 CT Main Auditorium
Track A Security Lead
TRACK A LEAD

Track A Security Lead

Announcing Soon

Gateways & Guardrails

Gateways & Guardrails: Mitigating Injection & Tool Poisoning

Defense-in-depth against indirect prompt injection, confused deputy attacks, real-time PII redaction, and semantic policy proxies.

02:00 – 02:45 CT Main Auditorium
Track B Lab 1 Facilitator
LAB 1 LEAD

Track B Lab Facilitator

Announcing Soon

Model Context Protocol Lead

Hands-On Lab 1: Production Model Context Protocol (MCP)

Facilitating live code construction of enterprise MCP servers with Server-Sent Events (SSE), scoped OAuth, and input validation.

01:15 – 02:45 CT Hands-On Lab
Track B Lab 2 Facilitator
LAB 2 LEAD

Track B A2A Facilitator

Announcing Soon

Distributed Multi-Agent Systems

Hands-On Lab 2: Multi-Agent A2A Interoperability & Trace IDs

Guiding engineers through Agent Cards (/.well-known/agent.json), asynchronous task states, and federated peer delegation.

03:00 – 04:30 CT Hands-On Lab
Community Participation

Have a production multi-agent case study?

We welcome technical lightning talks (10 min) showcasing real-world lessons learned, sandbox benchmark data, or open-source tooling.

Submit Speaker Proposal
Section 6

Ecosystem Partners & Sponsors

Supported by leading industry innovators, open-source consortiums, and engineering communities committed to advancing autonomous AI systems.

Community & Engineering Partners
IEEE Dallas Section

Advancing technology for humanity through joint community technical workshops, student mentorship, and engineering symposia.

Technical Co-Sponsor
AAIF Global

Autonomous AI Foundation providing open-source research frameworks, agent benchmarking suites, and community guidelines.

Parent Foundation
UTD AI Society

University of Texas at Dallas premier student AI organization driving machine learning research, agentic hackathons, and projects.

Student AI Organization
IEEE UTD Chapter

UT Dallas IEEE Student Branch empowering next-generation engineers through technical workshops and community engagement.

IEEE Student Branch
IEEE SMU Chapter

Southern Methodist University IEEE Student Branch fostering innovation, technical mentorship, and engineering excellence.

IEEE Student Branch
Event Sponsorship Opportunities
Direct Architect Collaboration & Hospitality Partnerships
Hospitality
FOOD & HOSPITALITY SPONSOR

Breakfast, Lunch & Breaks

Directly fuel 400 systems architects, engineers, and tech leaders across the full-day symposium.

  • Breakfast & artisan morning coffee service
  • Full catered hot networking lunch
  • Afternoon energy snacks & continuous refreshes
  • Dining hall banner & table branding
Openings for Breakfast, Lunch & Snack Sponsors
Official Swag
AAIF MEMBER PRODUCTS

Apparel & Merch Sponsor

Sponsor official AAIF member apparel and products directly from the AAIF Store for all 400 attendees.

  • Official AAIF member hoodies & tech tees
  • Co-branded attendee bags & badges
  • Store catalog: merch.aaif.io/collections/all
  • High-visibility lasting delegate presence
Official Merch Partnership Open
TECHNICAL LAB CO-SPONSOR

Sandbox & Compute Tooling

Provide cloud compute credits, model API access, and sandbox runtime environments for hands-on MCP and A2A code labs.

  • Pre-installed tooling in lab devcontainers
  • Sandbox API gateway credits for attendees
  • Prominent attribution in GitHub lab guides
  • Direct architect collaboration in live code
Limited to 2 Platform Partners
JOIN US AS AN INDUSTRY SPONSOR

Showcase Your Platform to 400+ Enterprise AI Architects & Engineering Leaders

Connect directly with staff engineers, systems architects, and technical decision-makers deploying mission-critical multi-agent systems across North Texas and beyond.