Dylan Dotolo. Leading product design at LiveKit.Setting the quality bar for AI-native product design, from first idea to shipped product.

Snapshots

Quick snapshots of the work I’m shipping each week, and the experiments I’m exploring along the way.

Voice agent evals

Designing a way to turn voice agent behavior into checks that teams can score and review.

Type:: Product design

Creating a greeting eval with a turn-count check — Creating an eval

Switching scoring methods in a criteria card — Configuring a criterion

I explored an eval builder that keeps each criterion’s question, scoring method, and failure behavior together. The design uses a yes-or-no check for a greeting and session data for a turn-count check.

Synthesis Design System

A design system that unified AWS AI/ML experiences built outside the console.

Type:: Design system

Synthesis buttons stacking into primary, secondary, tertiary, danger, hover, focused, and loading states

Synthesis started in my work on Amazon Bedrock Studio. As lead product designer, I grew it into a system other AWS service teams could use to stay on brand while building the experiences they needed. It eventually had its own team of designers and engineers.

The Bedrock Studio app shell with empty states for apps and components — The Bedrock Studio app shell

The Synthesis button documentation page showing button types — Button documentation

Project Howser

Guided onboarding that gets people using AWS AI/ML services in under a day.

Type:: Product design

The Project Howser intro reel moving through onboarding screens for SageMaker Canvas and Bedrock

Project Howser came out of the Machine Learning Thought Leadership team at AWS. As lead product designer, I worked with three service teams to redesign onboarding for Amazon Bedrock, SageMaker Canvas, and Amazon Transcribe. Adoption rose 30–60%, and time to value fell to 24 hours.

A grid of guided onboarding screens for AWS AI services — Onboarding across services

A SageMaker Canvas onboarding step asking how to improve the dataset — Improving a model in SageMaker Canvas

What If_

Exploring what the future of IBM Cloud could be.

Type:: Design strategy

The What If_ overview film: people working with concept IBM Cloud tools across devices and locations

What If_ is a design program I co-created with my manager to explore future IBM Cloud experiences. More than 200 IBMers took part. Together we set a north star for IBM Cloud, using storytelling to push innovation for customers.

A shared terminal where teammates and Watson set up a location — Collaborating in the terminal

An IBM Cloud concept feed with priority alerts and project news — The project activity feed

I’m a product designer based in Austin, currently leading product design at LiveKit. Previously, I worked at AWS, IBM, and Snowflake, designing developer tools and the systems behind them.

I like staying close to what gets shipped. That means setting direction with a team, working through a difficult flow, or building a prototype to see whether an idea holds up. I’m especially interested in complex products where a clearer interaction can help someone understand what’s possible and feel confident taking the next step.

My experience spans individual interactions, onboarding across multiple services, and design systems shared by teams. Across those scales, I care about making thoughtful decisions—and carrying them through to the details people actually use.

Outside of design, I enjoy traveling and exploring new cities, with time outdoors whenever I can.