Skip to contentTanner WilburnDrag to move Lucille. Click or press Enter to chat. Arrow keys move her; Escape returns her to the scroll path.
WASHINGTON, D.C.
Tanner Wilburn
Attorney at Mayer Brown. AI, cybersecurity, privacy & national security.
BACKGROUND
Before law, I worked in cybersecurity, with a focus on medical devices and radio frequency systems.
I still spend a lot of time building things. The projects below are personal work: ways to navigate a case record, handle routine tasks, and make research easier to inspect.
PERSONAL WORK
Projects
Tools for legal practice and security research.
LUCILLE / TASK ROUTINGILLUSTRATION
VOICE OR TEXT
A request
ON DEVICE
Timers Reminders Desktop actions
CLOUD
Research Reasoning Connected tools
A response on the desktop ↵
This diagram shows the desktop architecture. Click Lucille anywhere on the page to chat.
PROJECT 01 VOICE / LOCAL TOOLS / AI
Lucille
A desktop assistant.
Lucille runs on a Mac mini. She combines voice, reminders, research, and desktop actions, routing requests between local tools and cloud models.
Her eyes show when she’s listening, thinking, or responding. Here, she moves with the page.
“The agreement required delivery by Friday. The goods arrived the following Monday.”
Click a position to follow it back to its source. Fictional text only.
PROJECT 02 LAW / ARGUMENTS / DOCUMENTS
Argumentative.
Litigation analysis across documents.
A litigation workspace that connects claims, responses, evidence, and rulings across filings. Follow a thread through the documents, inspect the source, and build a response with the record in view.
Source passages stay connected to the working analysis.
Public sample available · sign in for your workspace.
JEV / REVIEW THRESHOLDSYNTHETIC EXAMPLE
Confidence and review
Party name98%Passes threshold
Defined term92%Passes threshold
Renewal date78%Review
Cross-reference61%Review
2 of 4 results sent for review.
Illustrative scores, not model outputs. A threshold is not a measure of correctness.
PROJECT 03 DECISION MODELS / EVALUATION
Jev
Experiments with decision models.
My experiments with TypeSafe’s Jev decision model explore how to turn qualitative questions into structured results—and how to test those results before relying on them in a workflow.
