I make AI products feel human.
Every great product starts with people. AI earns its place when it makes their day a little easier and a little more delightful.
How I work
The questions I get asked most, answered.
Start with the problem worth solving
I talk to users to understand why, use data to size it, then frame the problem so the team and leadership agree it's worth solving before we build.
Go wide, then bet on what matters
I explore lots of ideas with design and research, then narrow to the bets with the biggest impact and a clear way to measure success.
Prototype before the PRD
I build it first in Lovable or Claude Code, get feedback and run quick experiments, then write requirements from what worked.
Align everyone on the why
I bring clarity to ambiguous problems, then explain the why in each audience's language, from engineers and designers to data scientists and execs, so everyone moves in the same direction.
Guardrails from day one
I write evals with data scientists to check AI against human judgment, and work with legal, privacy, and safety teams on things like age limits before anything ships.
Plan for the worst, build for the best
Misuse, failures, and edge cases get planned up front, so the launch can aim high. I build a team culture where anyone can flag a risk. When one could change the product or timeline, I bring it to leadership early, with a plan, so there are no surprises.
AI is my sidekick. I call the shots.
AI keeps me scrappy. It's my thought partner, PRD co-writer, and meeting summarizer, and it turns my roadmap thinking into outlines and visuals I can share. That frees me to spend more time building with my team and delivering solutions.
Four products, start to finish
I'm grateful to have worked on so many different products. These four were the most pivotal. Each one walks through the problem, my role, key decisions, and results.
Can AI measure the skills a traditional test can't?
Exploring how AI can measure durable skills like communication, from early prototype to field pilot.
Read the case study →Taking Bing Image Creator from one model to three
Adding GPT-4o and Microsoft's first in-house image model without making creation confusing.
Read the case study →Launching Bing Video Creator
The first free way to use OpenAI's Sora, funded through a Microsoft Rewards credit strategy and launched mobile first.
Read the case study →From Flipgrid to Flip
A product rebrand and rebuilt app to take a classroom favorite further. +18% active users.
Read the case study →From Flipgrid to Flip
Flipgrid was a video tool educators and students loved, giving every learner a way to share their voice. I led the product side of a full rebrand, with a rebuilt app and a new name, so it could reach beyond the classroom.
The challenge
When Microsoft acquired Flipgrid in 2018, it was used by more than 20 million students and educators. The product and the name were built around one use: teacher-led video discussions in class.
People were already using it in more ways than that. To reach them, Flip needed a broader brand and a web app that was faster, simpler, and safer.
Under the hood, it was just as big a lift. The web app needed new routing and hosting, and the mobile apps needed a full rebuild.
My role
I led the full product lifecycle, from requirements to launch, across 6 engineering and design teams, and owned release readiness with support and community teams.
Key decisions
Rebuild the app, not just the logo
We shipped the new name with a rebuilt experience on web and mobile, so the rebrand felt like a real upgrade. In 9 months, we gave web new routing and hosting, migrated it to Remix, and rebuilt the mobile apps. I led the teams through every one of those changes at once.
Make the product tell the story
I worked with marketing on the "community for everyone" story and made sure the product delivered on it, from the first sign-up to the features people used every day.
You asked, we listened (YAWL)
YAWL was a guiding principle for the team: our users' voices should shape the product. As we redesigned, we focused on their most common asks, like threaded comments, an AI topic creator, and video analytics.
Treat the launch as part of the product
Launch communications, help content, and support training were planned alongside the features, so existing users weren't left behind by the change.
Results
In the press
Flip's core features have since moved into Microsoft Teams for Education.
Launching Bing Video Creator
The first free way for anyone to make videos with OpenAI's Sora, built right into Bing.
The challenge
Sora was one of the most talked-about AI models, but it was only available on sora.com, and only to people paying for ChatGPT Plus or Pro. Bing wanted to give everyone a way to try AI video for free.
That raised hard questions. Video generation runs on expensive GPUs, so how do you fund a free product? Which limits and features do you start with when competitors are moving fast? And how do you keep it safe and legal in every market?
My role
I led the product strategy and owned the end-to-end creation experience, from prompting to limits and the media canvas. I set the vision for AI-generated content as a growth driver for Bing and won senior leadership sponsorship. Along the way, I mentored two PMs who owned individual creation flows while I kept the product cohesive, and worked across engineering teams in the US and China and a design team. Every week, I briefed executives on goal metrics, status, and risks.
Key decisions
Make the GPU costs work
I built the funding and Microsoft Rewards strategy that justified the GPU spend. Standard-speed creations stayed free, and people earned fast creations by completing Rewards challenges, giving them more reasons to come back to Bing.
Pick the right limits, fast
I chose which limits and features to ship with, weighing cost and quality against a fast-moving competitive timeline so we could be first to offer Sora for free.
Start on mobile
I made the Bing mobile app our first home to increase app stickiness, while managing separate development timelines for every other entry point and surface. Desktop and search followed.
One flow across every entry point
I designed a consistent experience, from prompt to finished video, across the Bing app, Bing Image Creator, Bing Search, and Copilot Search.
Safe and legal across markets
With legal, privacy, and responsible AI teams, we set age limits, handled trademark concerns, and met local requirements in each market.
Results
Taking Bing Image Creator from one model to three
I led creation as Bing added GPT-4o and Microsoft's first in-house image model, MAI-Image-1, next to DALL-E 3.
The challenge
Bing Image Creator launched in 2023 on OpenAI's DALL-E. As newer models arrived, each was better at different things, and Microsoft AI was building its own image model.
The challenge was to add new models without making creation confusing, keep the experience fast and safe for every model, and bring it to the places people already search.
My role
I led the product strategy and owned the end-to-end creation experience, from prompting and image upload to creation limits and the media canvas where results appear. I also briefed executives weekly on goal metrics, status, and risks for both launches.
Key decisions
Make choosing a model simple
A model picker puts each model's strength up front, so people can pick the right one without knowing anything about AI.
Design one flow across every entry point
I designed the end-to-end flow, from prompt to finished image, across four entry points (Bing Image Creator, Bing Search, Copilot Search, and the Bing app) on web, mobile web, and native apps.
Set limits that work for people and the business
I designed creation limits for each model, balancing a generous free experience with the real cost of running image models.
Clear every bar before launch
I coordinated model quality reviews, responsible AI sign-off, and legal sign-off with engineering leads for each model before it shipped.
Bring creation to where people search
With the Copilot and Bing Search teams, we integrated image creation directly into their products, so creation wasn't stuck on a separate site.
Results
Can AI measure the skills a traditional test can't?
Exploring how AI can help students show durable skills like communication, and taking the idea from early prototype to a field pilot.
The challenge
For decades, tests have measured how students do in core academic subjects like reading, writing, and math. But employers and educators increasingly say that alone doesn't show whether a student is ready for college, a career, or life after school.
What's missing are durable skills, often called soft skills, like communication and collaboration. Schools want to build and track them, but they're hard to measure fairly and consistently, let alone at scale.
So at College Board's GenAI Studio, the challenge was: could we use AI to assess a durable skill well enough to give students a meaningful, trusted indicator, and do it at scale?
My role
I lead a cross-functional pod of designers, engineers, data scientists, learning scientists, and psychometricians. I own the product strategy and roadmap, and take the work from discovery to pilot.
Key decisions
Define the skill
With learning scientists and psychometricians, we defined what we were measuring and what "good" looks like before building anything.
Find where AI can recreate the experience
We looked for the moments where students naturally show a skill, then explored how AI could recreate that experience in a way that scales.
Prototype fast and get feedback
I built early prototypes myself and gathered feedback quickly to decide what to keep, change, or cut.
Measure AI against humans
With data scientists and psychometricians, I set up evals to check how closely the AI's judgments matched human experts.
Put it in real people's hands
We took it into the field so students and educators could use it for real, and learned what works outside the lab.
Results
Where I've been
- Dec 2025 to nowCollege BoardSenior Director, Product · GenAI Studio
Lead product strategy and roadmap for AI-enabled learning and assessment with a cross-functional pod. Took an AI-powered assessment concept from early prototype to field pilot.
- 2023 to 2025Microsoft AISenior Product Manager · Bing Multimedia
Owned AI image and video creation in Bing: Bing Video Creator, GPT-4o and MAI-Image-1 in Bing Image Creator, and creation in Bing Search and Copilot Search. Redesigned mobile video playback, doubling daily users of the video experience and adding a new revenue stream through video ads. Mentored 2 PMs who owned individual creation flows while I owned the end-to-end experience, working with engineering teams in the US and China.
- 2021 to 2023Microsoft Web ExperiencesSenior Product Manager · Flip
Led the Flip rebrand and rebuild across web and mobile, with 6 teams in 9 months, growing active users 18%.
- 2019 to 2021Microsoft EDUProgram Manager · Flipgrid
Built Flipgrid's first Customer Success team, hiring 5 people and an offshore vendor team, through 10x product growth, and cut daily support tickets 67% with a new help center, automation, and in-app help.
- 2017 to 2019CludoManager of Customer Experience, promoted from CSM
Managed a team of 5 across the US and EU, keeping 97% customer retention on $3.1M in revenue.
- 2014 to 2017Siteimprove and LeadpagesCustomer onboarding and customer success
Managed 233 accounts and $2.7M in revenue with 98% retention at Siteimprove, then ran training webinars and helped build a customer certification program at Leadpages.
- 2013College of Saint BenedictBA, Business Management
What I'm good at
Tools I use
A little about me
As a first-generation American and the first in my family to graduate from high school and college, I didn't have a playbook to follow. So I learned early that I get to design my own future.
That's how I approached tech, too. Everything I know, I picked up through my own curiosity or on the job: building with AI, shipping native mobile and web apps, and designing consumer-grade UX. I take a lot of pride in how quickly I learn, and it helps me navigate ambiguity and change with ease.
Being a mom of three kids under five has made me a stronger product manager. It's taught me to prioritize what really matters, make decisions quickly, and always advocate for the user (even the tiny, very loud ones).
I love working with people who love to build. That's where I do my best work. I've gotten to do it at every stage of the product development lifecycle (PDLC), from shaping product strategy and testing whether an idea is worth building in R&D, to launching new products and solving problems at scale.
Let's connect.
I love meeting people who are building products with AI that feel good to use. If that's what your team is working on, say hi on LinkedIn!