Reduces data complexity and costs by simplifying data architecture
Provides more accurate, personalized fitness recommendations to users
Improves user experience with more responsive, engaging interface
PEAR Health Labs consolidated its data and AI infrastructure with Google Cloud, using the combined power of BigQuery, Looker, and Vertex AI to deliver more personalized fitness recommendations to users.
PEAR Health Labs consolidated its data and AI infrastructure with Google Cloud, using the combined power of BigQuery, Looker, and Vertex AI to deliver more personalized fitness recommendations to users.
PEAR Health Labs is on a journey to use data and AI to help people achieve their health and wellness goals. Its platform includes more than 9,500 on-demand fitness sessions and supports a network of over 27,000 gyms and studios. This is powered by advanced PEAR Training Intelligence, which uses data and biometric feedback to build personalized fitness plans that drive real health outcomes.
To fulfill that promise, PEAR Health Labs relies on strategic technology partnerships. However, as the company has grown through acquisitions, it ended up with a fragmented tech stack and significant data sprawl across several cloud providers over the years, which made it difficult to scale its products and protect its customers’ sensitive health information.
“We faced escalating costs associated with our existing data warehousing and AI solutions, as well as tech redundancies,” says Jiwan Rai, vice president of data and AI engineering at PEAR Health Labs. “So we decided to consolidate our data and AI infrastructure with Google Cloud to achieve a new, AI-driven vision for the fitness industry.”
Here’s how migrating all of its data to BigQuery has empowered PEAR Health Labs’ primarily B2B customers with data- and AI-driven insights to support users’ holistic health journeys, and how this new architecture supports PEAR Training Intelligence.
We faced escalating costs associated with our existing data warehousing and AI solutions, as well as tech redundancies. So we decided to consolidate our data and AI infrastructure with Google Cloud to achieve a new, AI-driven vision for the fitness industry.
Jiwan Rai
VP of Data and AI Engineering, PEAR Health Labs
Adopting a simplified and unified data architecture
PEAR Health Labs has a long history with Google Cloud and a lot of its core products were built on the platform. When the team evaluated unified data platforms, it looked beyond pure warehousing functionality to find a solution that would also support seamless data integration, AI innovation, and data governance needs. Ultimately, consolidating on BigQuery would provide the best financial, operational, and long term-technical value.
Google Cloud is also highly committed to security, with robust features that would bolster our data security posture. These built-in tools — such as role-based access control, data loss protection, and classification of sensitive data — would provide everything PEAR Health Labs needed to meet data security and compliance requirements without getting in the way of innovation.
Working with the Google Cloud team, PEAR Health Labs began a multi-stage migration to BigQuery, running both its previous and new solutions in parallel to minimize disruption. “One of the biggest challenges was migrating complex business logic that had built up over many years across multiple business units, including those brought in through acquisitions,” Rai explains. “Adapting to the different paradigms and SQL functions between the two systems also required a learning curve for our engineering team, but once we mapped our architecture in the new environment, we were able to move very quickly.”
Now, by adopting a tiered data transformation model within BigQuery, PEAR Health Labs has significantly simplified its data architecture.
Adapting to the different paradigms and SQL functions between the two systems also required a learning curve for our engineering team, but once we mapped our architecture in the new environment, we were able to move very quickly.
Jiwan Rai
VP of Data and AI Engineering, PEAR Health Labs
It directly ingests raw data sources into BigQuery, transforms them into standardized business entities, and then organizes them into bronze, silver, and gold layers for refinement. This streamlined approach provides several benefits: clear data lineage, reduced complexity, and cost savings by materializing one set of core entities instead of many. This simplified data architecture empowers the team to deliver accurate and timely insights to drive better business decisions and provide stronger recommendations through PEAR Training Intelligence to help users make better health decisions.
Activating real-time data processing to encourage healthier lifestyles
With its data infrastructure centralized in BigQuery, PEAR Health Labs can now perform holistic analysis across projects in one powerful platform. As the company has moved into the wearables space, it has the flexibility to stream real-time, event-driven data such as user activity. Through its direct integration with Pub/Sub and Cloud Storage, the team can ingest and process this data as it arrives, rather than relying on batch processing.
Aaptiv AI can help users navigate our extensive content library by understanding user queries and preferences and recommending tailored workouts and programs. For instance, a user might ask, ‘What's a good workout for a beginner?’, and Aaptiv AI will suggest relevant workouts.
Jiwan Rai
VP of Data and AI Engineering, PEAR Health Labs
This real-time capability is crucial to PEAR Training Intelligence’s ability to provide actionable insights to enterprise customers. For example, when a Fitbit user completes an activity (such as going for a run) or records a wearable metric (like a nightly sleep score), PEAR Training Intelligence can immediately process that data and provide personalized health or fitness recommendations. Analyzing these insights enables a more dynamic and responsive user experience — but even more importantly, it helps identify crucial health metrics like how increased physical activity can lower risk of chronic disease.
To further enhance real-time data processing and improve its AI-powered applications like PEAR Training Intelligence, the company is using Vertex AI and running proofs of concepts with BigQuery ML (BQML).
The accessible, user-friendly interface in Vertex AI makes it easier for teams to experiment with AI and machine learning (ML) models, without requiring deep technical expertise, while the integration between BigQuery with Vertex AI allows them to directly leverage data to train and deploy models. The company can then visualize and analyze data in Looker, a business intelligence platform from Google Cloud, to identify opportunities to improve its products. “We’re excited about the potential of BQML to perform tasks like sentiment analysis on unstructured data, further powering innovation as we grow,” Rai says.
One of the most exciting developments with PEAR Training Intelligence is the addition of an AI-powered chatbot that can ingest Fitbit workout data that has been published to Apple HealthKit. This chatbot, named "Aaptiv AI," provides users with a natural language interface to interact with the platform.
“Aaptiv AI can help users navigate our extensive content library by understanding user queries and preferences and recommending tailored workouts and programs,” Rai shares. “For instance, a user might ask, ‘What's a good workout for a beginner?’, and Aaptiv AI will suggest relevant workouts. Or if a user is training for a marathon, the chatbot can recommend a progressive training plan.” This tool will enhance the user experience by making it easier to find the right workouts and stay motivated on their fitness journey.
By fully migrating to BigQuery, PEAR Health Labs has positioned itself to better leverage the power of data, AI, and ML. This unified platform enables the company to efficiently analyze vast datasets, extract valuable insights, and develop innovative solutions that cater to the evolving needs of its end users. As the team looks to the future, personalization will remain at the core of its mission. With the enhanced capabilities of BigQuery, PEAR Health Labs can go beyond simple data analysis to empower individuals with actionable insights and drive healthier lifestyle choices.
“Through our strong partnership with Google Cloud, we're committed to shaping a future where healthcare is more personalized, preventive, and effective,” Rai says.
PEAR Health Labs is a fitness technology company that builds hyperpersonalized coaching solutions to help people move smarter, get healthier, and live happier.
Industries: Retail and Consumer Goods
Location: United States
Products: BigQuery, Cloud Storage, Looker, Pub/Sub, Vertex AI
