Speaking
I speak on the systems I build and the work I have shipped: multi-agent AI architecture, what AI assistance does to the person receiving it, AI adoption across a large engineering organization, and autonomous trucks inside a working fleet. Below are the four subjects I can take to depth, where I have spoken, a bio you can paste into a program, and how to book me.
What I speak on
Four subjects, each one built on work I have shipped. Every session is shaped for the room it is in, so read these as the territory a session can cover.
Designing multi-agent systems that check their own work
The Orchestrator-Worker-Antagonist pattern, and what it costs to run. A worker agent produces, an antagonist agent argues against the output from a written standard, and an orchestrator makes the final call, with each role on a different model family. I cover what an adversarial reviewer inside the loop changes, the parts nobody puts in the diagram, and the cases where the pattern is not worth paying for. The full specification is on the OWA reference page.
What it draws on
- Ember runs this architecture in private alpha across Claude, GPT and Gemini, connected to 30+ external services over MCP
- The same architecture makes the scaffolding decisions in my thesis instrument, where every decision records the proposal, the objection, the rationale and which model served each role
- Failure handling is designed up front: per-role fallback models, a deterministic fallback marked as such so those cases can be analyzed separately, and an incident row naming the agent, the error type and the attempt
Engineering and AI practitioners who want the mechanics in detail.
Measuring what AI help does to the person receiving it
Cognitive load theory used as a product instrument. Where intrinsic, extraneous and germane load come from, why fading support too fast and too slow fail in different ways, and how to design a controlled study a product team could run before it ships an assistive feature.
What it draws on
- My MS thesis at the University of Nebraska Omaha: a two-condition study putting AI-driven adaptive scaffolding against a fixed faded schedule of worked examples
- Instrumented end to end, from consent through pretest, learning activity, an eight-item Klepsch cognitive load assessment and posttest, with server-side validation on every response
- Protocol, instruments and the multi-agent scaffolding system are built and awaiting IRB approval
HCI and education technology researchers, and product teams shipping assistive AI features.
Getting an engineering organization to actually adopt AI tooling
What changes when AI coding tools land in a large engineering organization, and what a licence count tells you (very little). Enablement that reaches people who did not ask for it, the evaluation practice a team needs before it trusts a model in its workflow, and the difference between a pilot and a habit.
What it draws on
- Helped onboard 300+ developers, product managers and IT staff to AI coding and productivity tools
- Shaped AI-enablement strategy with technology leads across 20+ product engineering teams company-wide, a broader remit than the teams I owned product direction for
- Led product direction for 13 agile teams and roughly 150 engineers, through a product and UX team of seven who each owned strategy for their own teams
Engineering and product leadership.
Putting autonomous trucks inside the systems a fleet already runs
Autonomy treated as an integration problem. An autonomous truck handled as an ordinary fleet asset inside the transportation management system a carrier already operates, across load tendering, shipment updates, exception events and maintenance, and what it takes to get a frontier technology partner and a 24/7 operation onto the same roadmap.
What it draws on
- Partnerships and integrations with Kodiak, Aurora, TuSimple, Embark, Daimler, International and PACCAR
- First fleet in the US with direct, bidirectional integrations to every major Class 8 OEM maintenance system
- Platforms consolidating 20M+ events a day across a 7,600-truck, 30,000-trailer operation
- Kodiak published its own account of the integration program, and the work was covered by Transport Topics and Samsung
Fleet operators, autonomy teams, and anyone putting frontier technology inside an operating business.
Formats
What I will turn up and do. Length, depth and the amount of live demo are all adjustable; tell me the audience and I will tell you what fits.
- Conference session
- Workshop for a practitioner audience
- Podcast or recorded interview
- University guest lecture
- Internal leadership or all-hands session
- Panel
Where I have spoken
Guest lectures at the University of Nebraska Omaha, and Werner's own events. Product Days was Werner's annual product event, presented to hundreds of employees, vendors and customers; the Innovation Series was the company podcast.
Teaching
- Guest Lecture: IoT, Big Data, and the Cloud in TransportationUniversity of Nebraska Omaha, 2024
- Guest Lecture: IoT and Cloud Platforms for TransportationUniversity of Nebraska Omaha, 2022
Werner Enterprises
- Werner Product Days 2024Werner Enterprises, 2024
- Werner Product Days 2023Werner Enterprises, 2023
- Innovation Series Podcast: Autonomous TruckingWerner Innovation Series, 2023
- Werner Product Days 2022Werner Enterprises, 2022
- Innovation Series Podcast: Big Data in Fleet OperationsWerner Innovation Series, 2022
- Innovation Series Podcast: Gate AutomationWerner Innovation Series, 2022
- Innovation Series Podcast: Fleet Maintenance TechnologyWerner Innovation Series, 2021
Each of these is written up on the media and speaking page, alongside the writing and the awards.
Bio
Two lengths, written in the third person so they can go straight into a program or an event page. Copy either one; no attribution needed.
Short bio (53 words)
Tristan J. Nolan is a Sr. AI Platform Product Manager and an MS candidate in Human Centered Computing at the University of Nebraska Omaha. He designs and builds multi-agent AI systems and researches how AI should scaffold human learning, on top of 12+ years at Werner Enterprises delivering enterprise transportation platforms at scale.
Long bio (199 words)
Tristan J. Nolan is a Sr. AI Platform Product Manager and an applied AI researcher. He designs and builds multi-agent systems, including Ember, which runs an Orchestrator-Worker-Antagonist architecture where an adversarial agent evaluates every output before it is released. His MS thesis at the University of Nebraska Omaha, completing December 2026, measures AI-driven adaptive scaffolding against a fixed faded schedule of worked examples and what each does to a learner's cognitive load and skill acquisition.
Before that he spent 12+ years at Werner Enterprises, where he led product direction for 13 agile teams and roughly 150 engineers, delivered platforms processing 20M+ events a day, built the first US fleet with direct bidirectional integrations to every major Class 8 OEM maintenance system, and partnered with Kodiak, Aurora, TuSimple and Embark on autonomous truck integration. He helped onboard 300+ developers, product managers and IT staff to AI coding tools, shaped AI-enablement strategy with technology leads across 20+ product engineering teams, and guest lectures at the University of Nebraska on IoT, big data, cloud infrastructure and AI adoption. He was named one of Heavy Duty Trucking's Emerging Leaders in 2020 and a finalist for Samsara's Technology Leader of the Year in 2023.
Booking
Send the date, the audience, the format and the length, plus anything you already know about what the room needs to walk away with. I answer within two to three business days.
Book a talkGuest lecturing is a standing yes. If you teach an HCI, software engineering or transportation technology course and want a practitioner in the room, just ask.
Common questions
What is Tristan J. Nolan available for?
Tristan J. Nolan is available for speaking engagements, advisory conversations, and consulting on AI strategy and product, agentic implementations, product engineering, autonomous systems, and fleet technology. The speaking page carries the subjects, formats and bios; the contact form is the way in.