Talks and Appearances
Talks and appearances.
More than a decade in data science and AI. I lead enterprise AI delivery at a regulated insurer, teach Python and machine learning, and build products myself. The rooms I speak to usually hold engineers and executives at the same time, and I write for both without diluting either one.
Book me to speak →Featured talks
- Prairie Python, hosted by ImprovingOmaha · September 2026
Agents Aren't You: Building Python Repositories for First-Class Agent Identity
Agents are already writing code in Python repositories, and their pull requests merge at close to the human rate. The question stops being whether to let them in and becomes what the repository tells them when they arrive. My answer is skills: a folder with a SKILL.md that an agent opens only when the work calls for it, so knowledge only your codebase can supply is cheap to keep and paid for only when it is used. I showed the pattern running in Plasmid, the autonomous Python crew that maintains this site without me in the loop.
The part I care about most is identity. NIST frames an agent as an identity in your system rather than a feature of an app: its own credentials, scoped entitlements, and approval that means something instead of consent fatigue. That is a design decision available to us right now, and a retrofit if we wait.
- Women in Technology of the Heartland (WiTH)Omaha · August 2026
The Human Side of AI Software Engineering
What changes for technical teams when AI is in the stack, and what gets harder rather than easier. Product judgment, experimentation, data quality, and evaluation all matter more than they did, and accessibility stops being something you add at the end.
The working example was Selvo, the personalized communication companion I am building for nonspeaking adults: what happens when a system is built to adapt to a person instead of requiring the person to adapt to it. It is also where I keep landing on the same question. How do we help technology understand a person without letting it silently speak for them?
WiTH Omaha · August 2026 - Databricks Data + AI World TourChicago · 2025
Financial Services Forum
How a new AI organization got two products into production inside a regulated insurer, presented with colleagues from our data platform and architecture teams. Smart Claims Assistant and Underwriting Assistant, the Databricks foundation underneath them, and the governance that made both defensible.
The through-line was augmentation. People contact us on one of the worst days of their life, so automating humans out of that was never the goal. Pair expert people with capable machines and get the customer a decision sooner.
I also covered how we govern the work, built on NIST, NAIC, and LIMRA responsible AI principles, and why KPIs have to be designed per workflow. A model being right is not the same as operations capturing value.
Databricks Data + AI World Tour · Chicago, 2025 - Heartland Developers Conference and Big AI, AIM InstituteOmaha · 2024
Keynote AI Leaders Panel
Panelist on value realization in AI delivery, and on governance by design: building the controls in from the start so they accelerate delivery rather than gate it at the end.
The panel put four AI leaders from different industries next to each other, which surfaced how much of AI leadership is really leadership under uncertainty. How do you justify investment before you can prove ROI. How do you keep an engineering team confident while models change under their feet. How do you answer a board that asks whether a system is safe, when the honest answer is "safe for what."
The version I keep coming back to is that governance is the shape of a program, not the paperwork attached to it. If the controls have to be added at the end, they are compliance theater. If they are how the team makes decisions from day one, they are the reason the program can move quickly enough to be worth doing.
Keynote AI Leaders Panel · Heartland Developers Conference, 2024