WHOOP Innovation & Technology Culture

What practices does your team employ to foster innovation, and how have these practices led to more creative, out-of-the-box thinking?

I am on the AI support team and was previously on the AI coach team that empowered users to chat with all of their data with AI. This space moves incredibly fast, so everyone on the team is constantly reading what models are being released, how costs are changing and how we can pull in user data more effectively to better answer members’ questions. We have a repository explicitly for doing experiments with AI that has been incredibly useful. 

In this repository, we constantly run small tests against our existing code and new experiments we want to try and measure the results to be metric-driven. This results in trying new retrieval patterns, new models and new strategies for projects. It also often means making things that are extremely simple with large language models and then rapidly iterating as often as we can to ship big features very quickly without having to define a lot of logic to support them, as the LLMs can work through the edge cases on their own.

How has a focus on innovation increased the quality of your team’s work?

We are currently working on classifying languages and automatically responding to support cases in the correct language. By focusing on innovation, we already have the infrastructure to test a variety of different approaches very quickly. We then watch metrics move and adjust accordingly. Language classification is surprisingly tricky to do well, and LLMs will only get it right about 90 percent of the time, which would result in a terrible member experience. By pulling in a different library, deploying it and then iterating, our team was able to make this 99.9 percent accurate. This was only possible due to our constant focus on building metrics, looking for better tooling and being data-driven.

Some other examples include pulling in external benchmarks when determining the best AI models to use in our code assistance tools, AI support or coaching system, as well as building our own benchmarks and using Anthropic’s model context protocol to accelerate our development team or our products.

How has a focus on innovation bolstered your team’s culture?

By focusing on innovation, we are constantly all learning, experimenting and lifting each other up. We all are learning new tools, and when we see metrics go up, we celebrate together. We make sure we teach each other what works and quietly let go of the things that do not. Building at the bleeding edge is a lot of fun, and we are building the infrastructure to enable that.

What’s your professional and academic background, and how did you break into the tech industry?

My path into tech wasn’t traditional. I studied business and film at Brooklyn College, planning for a career in entertainment. After earning my undergraduate degree, I moved to Los Angeles and joined a talent agency, Creative Artists Agency, in their corporate communications department, where I worked on public relations for both the agency and its clients. After CAA, I moved to ITV America, which is a non-scripted television production company. This was a super exciting time. I supported the development and production of shows like Queer Eye and Love Island while also helping independent producers bring new show ideas to life.

Eventually, I felt burned out by Hollywood and wanted to return to Boston, my hometown. I joined Accomplice, a venture capital firm. I quickly fell in love with supporting early-stage founders building startups and fellow tech operators. Instead of working with independent producers on pitching TV shows, I was supporting early-stage founders building startups. 

That role marked my official entry into tech. I met my current manager, Ryan Durkin, through a partner there, and we’ve been on a journey together across three startups: The Operators, AnyQuestion and now WHOOP, as a result of its acquisition of AnyQuestion.

How did you learn how to use AI, and how do you apply it to your work? 

I started using AI regularly in 2023 while working at AnyQuestion. We were building the “Substack of Knowledge,” a question-and-answer platform featuring experts like Olympic athletes and top doctors answering AI-generated questions. I used AI to generate social content and help our creators organize their voice-to-text responses. That’s when I realized how powerful AI could be. WHOOP’s AI @ Work team is focused on enabling an AI-powered workforce. We ensure our team has access to the best AI tools, learns how to use them and feels supported. People come to us with ideas and problems, and we aim to help them solve them with AI. 

AI can be intimidating. We want people to feel safe and comfortable experimenting. I learn mostly from colleagues and experimentation. Our companywide AI Slack channel is full of shared resources and clever use cases. We also host internal learning and development sessions covering everything from prompt writing to tool-specific tutorials. Outside of work, I’m constantly testing new AI tools to see what we may want to adopt. Lately, I’ve been into video generation platforms like Kling, Runway and Veo, which we believe will be powerful for teams like marketing and industrial design.

What do you consider the greatest benefits of leveraging AI, and how has it positively impacted your career?

AI isn’t just about saving time; it’s a force multiplier. It helps us develop skills and make better decisions. My favorite use case is thought-partnering. When I’m stuck or unsure about how to start a project, I open ChatGPT, use voice-to-text, and just talk it out. Then I go back and forth with it as I would with a teammate in order to find the right direction. That process has helped me trust my judgment more. I’m also building skills I never imagined: I can vibe code simple apps, create professional designs, and even generate animated films. That kind of growth used to take months or years — now it’s at my fingertips.

To wrap things up: Always remember that AI is a tool. Your work is still your work, and you should be proud of what you put out into the world. Did I use ChatGPT to help review and edit these responses? Yes. It’d be silly of me not to. Did I write everything from the heart, take time to draft a precise prompt to ChatGPT to ensure it was reviewed and edited as I wanted, read and reread to make sure no errors were made, and my voice still came through? Also, yes. AI’s evolving daily, but so am I. That’s what makes this ride so exciting.

What’s your rule for fast, safe releases — and what KPI proves it works?

I’ve found that speed comes from ownership. We decentralize so that engineers own domains, not tasks, and that ownership doesn’t end when code merges. You own the outcome, which means following through to make sure what you shipped actually works the way you intended.

We start projects with a tech plan to get alignment and catch bad ideas early. Then we break work into two-week chunks maximum. If it takes longer, you haven’t split it up enough. The KPI that matters? User metrics. Everything else (PRs shipped, A/B tests launched) is just a leading indicator.

What standard or metric defines “quality” in your stack?

With AI, errors don’t look like traditional errors. The failure modes are much more subtle and hard to track down. A bad response doesn’t throw an exception, it just gets ignored. Users quietly lose trust and you can’t trace that in a log file. So we built a system where AI grades our AI.

We have a set of “golden questions” that cover the range of things members actually ask. Automated agents grade responses on personalization, refusal rate, correctness. We validated these agents against human reviewers to make sure their judgment holds up. We also run this on live production traffic, not just test environments, so we catch regressions before they compound.

Name one AI/automation that shipped recently and its impact on your team or the business.

I’m most proud of the Advanced Labs Uploads. Members can upload bloodwork from any lab as a PDF or screenshot, in any language, and we extract the biomarkers automatically. Before writing production code, we built an eval set from manually graded reports. Tested different orchestration approaches, iterated on prompts until accuracy was solid. That eval work upfront is what lets us ship without second-guessing ourselves.

The feature has been a great success. We’ve seen a ton of uploads since launch and members can now see their lab results alongside their WHOOP data and get coaching from WHOOP Coach on what it all means.