Get in touch

AI at Amazon

Partnering with Amazon Bedrock to make hiring smarter.

Partnering with Amazon Bedrock to make hiring smarter.

In mid-2021, I spearheaded a project with Amazon Bedrock, engaging in their first real-use program before public launch.

I led design, research, and strategy, facilitated cross-team collaborations and workshops, worked closely with data scientists, and explored new frontiers in behavioral science.

In mid-2021, I spearheaded a project with Amazon Bedrock, engaging in their first real-use program before public launch.

I led design, research, and strategy, facilitated cross-team collaborations and workshops, worked closely with data scientists, and explored new frontiers in behavioral science.

In mid-2021, I spearheaded a project with Amazon Bedrock, engaging in their first real-use program before public launch.

I led design, research, and strategy, facilitated cross-team collaborations and workshops, worked closely with data scientists, and explored new frontiers in behavioral science.

This initiative positioned us at the forefront of practical AI application and helped us build a truly intelligent experience that anticipates the needs of managers across Amazon.

This initiative positioned us at the forefront of practical AI application and helped us build a truly intelligent experience that anticipates the needs of managers across Amazon.

This initiative positioned us at the forefront of practical AI application and helped us build a truly intelligent experience that anticipates the needs of managers across Amazon.

2021-2023

2021-2023

Year

Disciplines

Human interface design

Interaction design

Design systems

Research

Product strategy

Art direction

Preview

Preview

Predictive insights and smart recommendations based on real-time events, ensuring managers are always moving forward with their hiring goals.

Machine learning

Natural language processing

Generative AI

Do less to accomplish more

Do less to

accomplish more

Naturally surfaced shortcuts, to help managers accomplish more by doing less.

We use machine learning to identify matching open positions, auto-selecting them so they don't have to.

Machine learning

Smart questions for

smart decisions

Smart questions

for smart decisions

Smart questions

for smart decisions

Search that anticipates needs, offers actionable insights, and facilitates effective decision-making.

Machine learning

Natural language processing

Generative AI

What did I learn from this work?

What did I learn from this work?

What did I learn?

What did I learn?

Find shared language to co-create.

Find shared language to co-create.

I spent a lot of time working with data scientists to build AI models. when building worlds, we often explored similar experiments and assumptions.

They use technical intuition and I study behavior and systems. We communicate through diagrams of observations, entities, outcomes, and systems.

For every collaborator, there is a shared language; find it and find a key.

What did I learn?

What did I learn?

Share thoughtfully.

Share thoughtfully.

This work caused such excitement and potential, it was closely observed by executive leadership at Amazon. I was responsible for owning monthly progress review with these leaders. Each review was only 25 minutes long.


I relied on the following format to effectively lead each review:

What we want

We want to achieve this business outcome

How we think we’ll get it

If we seed this behavior (prototype) we will achieve the outcome

Why we think that

We can achieve the outcome through that behavior because of this evidence

When it should happen

Here is when we’ll deliver to users

How we’ll know we did it

Here is how we’ll know we’ve successfully reached that outcome

Most problems are search problems.

Most problems are search problems.

When we help users get to what they need, it’s through search. Navigating, way-finding, browsing, personalizing, recommending are all search problems.

Search doesn’t have to be the action of a human. An intelligent experience should search on behalf of them; surfacing anticipatory suggestions that appropriately drive behavior.

Articulation is a barrier.

Articulation is a barrier.

Humans simply struggle to articulate their way into getting the results they desire through prompting. It requires back and forth and long streams of text,
which inevitably impacts usability.

Instead, identify their motive through data and anticipate it in the experience.


Develop a bit of of contrarian thinking.

Develop a bit of of contrarian thinking.

Working with emerging and revolutionary technology can actualize powerful experiences, but it also means you what you create can create dramatic impacts on people, so you must embrace the responsibility of defending their rights.

AI is not a feature you need to market.

AI is not a feature you need to market.

In 2021, the trend of labeling every feature with "AI" was already on the rise, but I led us down a different path. I maintained our focus on the quality of the user experience rather than the underlying technology.

After all, the goal is to provide the best possible service, regardless of whether it involves machine learning, natural language processing, or any other form of AI.

In 2021, the trend of labeling every feature with "AI" was already on the rise, but I led us down a different path. I maintained our focus on the quality of the user experience rather than the underlying technology.

After all, the goal is to provide the best possible service, regardless of whether it involves machine learning, natural language processing, or any other form of AI.

This website is powered by HTML

What difference did we make?

What difference did we make?

What difference did we make?

What difference did we make?

Less than

Less than

11 min

using Roster each week

More than

More than

$50M

$50M

$50M

In time saved by managers

Where we started

We built Amazon Roster to make it natural to manage headcount and hire to plan.


I led research, product strategy, and design—and forged our strategic vision and product roadmap.

We built Amazon Roster to make it natural to manage headcount and hire to plan.


I led research, product strategy, and design—and forged our strategic vision and roadmap.

What difference did we make?

What difference did we make?

Less time

11 min

using Roster each week

More than

$50M

In time saved by managers

Twitter

Are.na

Read.cv

Twitter

Are.na

Read.cv

Twitter

Are.na

Read.cv

Twitter

Are.na

Read.cv

Where we started

We built Amazon Roster to make it natural to manage headcount and hire to plan.


I led research, product strategy, and design—and forged our strategic vision and product roadmap.

Where we started

We built Amazon Roster to make it natural to manage headcount and hire to plan.


I led research, product strategy, and design—and forged our strategic vision and roadmap.

Back to top