Speaking · Atlanta, Georgia
I speak about what has to be true before a person acts on what a system tells them. Seventeen years building decision-heavy products at Meta, Pinterest, Opendoor, and CNN.
Available for conferences, panels, and podcasts.
Every automated system asks a person to act on something they didn’t compute. What nearly two decades of platform research says has to be true before they will.
The answer a person is looking for is usually already in the system. The interface is what’s hiding it.
Conversational interfaces discard the shape of the thinking that produced them. That is where trust breaks.
Working sessions for product and design teams, delivered on site or remote.
When to act as a delivery engine and when to act as a co-pilot, and why the choice is situational rather than seniority-based. Includes a two-layer gate for deciding what matters when everything is important.
Delivered to NPR’s design team, June 2025.
Seventeen years across Meta, Pinterest, Opendoor, and CNN. Embedded product and design work across a Techstars Atlanta cohort. MBA in Applied Artificial Intelligence, Northeastern University. Founder of Bayesian Leaf, a fractional product and design practice.
Invited talk, NPR design team, June 2025.
Earlier talks during a startup run, 2014 to 2015.
Square, 1000 by 1000 pixels.
Hew Suber is a fractional product and design leader who works on AI products that decide, recommend, and act on a person’s behalf.
Hew Suber has spent seventeen years building decision-heavy product systems at Meta, Pinterest, Opendoor, and CNN. He now runs Bayesian Leaf, a fractional product and design practice for AI-native teams, and writes about what has to be true before a person acts on what a system tells them.
Hew Suber is the founder of Bayesian Leaf, a fractional product and design practice for AI-native companies. He has spent seventeen years building decision-heavy product systems, including automation, self-serve, and operational modeling, at Meta, Pinterest, Opendoor, and CNN, plus embedded work across a Techstars Atlanta cohort. At Pinterest he led design for an AI automation product built from nothing on a product line generating over $600 million annually. At Opendoor he mapped the operating platform for a national real estate business as it scaled from 21 markets to 51. His work centers on one question: what has to be true before a person acts on what a system tells them. He holds an MBA in Applied Artificial Intelligence from Northeastern University and writes at bayesianleaf.substack.com. He lives in Atlanta.
To invite me to speak: hew@bayesianleaf.co
