Use Case
Coinbase, a leading cryptocurrency platform, uses machine learning for critical operations including fraud detection, personalization, and blockchain data analysis. Processing user actions in milliseconds and managing thousands of features across hundreds of pipelines, Coinbase needed a robust solution to scale their ML infrastructure. This led them to leverage Tecton's feature platform.
consectetur adipiscing elit
consectetur adipiscing elit
consectetur adipiscing elit
When Coinbase first started with machine learning over eight years ago, their approach was entirely ad hoc, with teams building pipelines manually. While they attempted to build in-house solutions, they found feature engineering to be a common problem that was challenging to solve with a small team.
“What shines about Tecton is the feature engineering experience—that developer workflow,” explains Joseph McAllister, Senior Engineer on Coinbase’s machine learning platform team. “From the very beginning, when you’re onboarding a new data source and building a feature on Tecton, you’re working with production data, and that makes it really easy to rapidly iterate.”
The improved workflow has enabled more people to build streaming pipelines without requiring the same domain expertise as before, while individual engineers can iterate faster on their pipelines.
Prior to Tecton, feature development often required multiple teams and handoffs. “There is no concept of one person designing a feature and prototyping it, and a separate team productionizing and maintaining that feature,” McAllister notes.
Now with Tecton, a single person can develop features, generate training data, validate it, and deploy to production. Engineers can prototype features and test them with real production data without creating pull requests or waiting on other teams.
Engage with a product expert in a one-on-one session to explore what you can do with Tecton.
Our interactive demo allows you to explore and try our AI Data Platform in just 5 minutes!
For Coinbase’s fraud detection system, speed is critical. When a user takes an action, such as attempting to withdraw funds, the system needs to fetch features, generate a fraud prediction score, and take appropriate action—all within milliseconds.
“This is something that happens instantaneously, and happens every time a user takes an action at Coinbase,” explains McAllister. Tecton delivers the required feature values with very low latency, which is essential for these real-time use cases.
The impact of Tecton on Coinbase’s ML operations has been transformative. “Without Tecton, I don’t think we’re going to be able to get the most value out of machine learning. It would become a bottleneck around feature engineering,” McAllister states.
The platform has enabled Coinbase to:
Looking ahead, McAllister sees Tecton’s potential expanding: “Tecton already is very effective at transforming different kinds of data and generating datasets… There’s a lot that can be done around computing and serving embeddings, as well as getting other real-time context into models.”
By providing robust infrastructure for feature management, Tecton has become a robust addition to Coinbase’s ML operations, enabling them to focus on building features that drive business value.
Engage with a product expert in a one-on-one session to explore what you can do with Tecton.
Our interactive demo allows you to explore and try our AI Data Platform in just 5 minutes!
Unfortunately, Tecton does not currently support these clouds. We’ll make sure to let you know when this changes!
However, we are currently looking to interview members of the machine learning community to learn more about current trends.
If you’d like to participate, please book a 30-min slot with us here and we’ll send you a $50 amazon gift card in appreciation for your time after the interview.
or
Interested in trying Tecton? Leave us your information below and we’ll be in touch.
Unfortunately, Tecton does not currently support these clouds. We’ll make sure to let you know when this changes!
However, we are currently looking to interview members of the machine learning community to learn more about current trends.
If you’d like to participate, please book a 30-min slot with us here and we’ll send you a $50 amazon gift card in appreciation for your time after the interview.
or
Interested in trying Tecton? Leave us your information below and we’ll be in touch.
Unfortunately, Tecton does not currently support these clouds. We’ll make sure to let you know when this changes!
However, we are currently looking to interview members of the machine learning community to learn more about current trends.
If you’d like to participate, please book a 30-min slot with us here and we’ll send you a $50 amazon gift card in appreciation for your time after the interview.
or