2021 at Tecton: A Year in Review

2021 is a wrap, and what a year it has been! As we head into the New Year, I want to take a look back at what we’ve accomplished this year and share with you some of the things I expect for 2022.

Tecton has experienced a huge amount of growth this year. The team has many new faces and we’ve adapted to working in hybrid mode with re-opened offices in New York and San Francisco (Seattle is coming soon!). 

Along the way, we’ve celebrated a large number of company and product milestones that support our growing customer base, which includes Fortune 100 insurance companies and banks, gaming giants, growing fintech firms, SaaS providers, e-commerce sites, health-tech startups, and many more.

The data industry exploded

It has been an incredible year for the data and Machine Learning industry. We predicted that 2021 would be the year of the Feature Store and, as it turns out, the space exploded and exceeded all of our expectations! Databricks, AWS, and many others promoted their own versions of a Feature Store. I’m thrilled to see so much innovation in this space, and so much attention paid to the data-centric view of ML, as it is steadily getting us closer to developing and productionizing Machine Learning applications with ease.

Now, of course, Tecton has benefited from the data industry as a whole building up a massive tidal wave. Snowflake, Databricks, Confluent, dbt, and several others have been going through the roof and modernizing enterprise data stacks. The Modern Data Stack has been one of the most talked-about paradigms in 2021. All of which is fantastic news for ML, as these companies are establishing the data foundation upon which Machine Learning applications are built.

A growing team

There is a classic saying, “Startups reinvent themselves every 6 months,” which could not have been more true: 

• The team of “Tectonauts” 🧑‍🚀 (a highly debated name!) nearly tripled and has grown into a diverse group of over 50 talented people from many backgrounds and previous experiences. To learn more about them, check out our “Featured Tectonaut” series, which started with with Samantha and David.
• We’ve eased into a new style of building hybrid office hubs in major cities while also growing our team of fully remote Tectonauts. We now work out of 7 different cities with more on the way!
• 7 Tectonauts welcomed a baby to their family 🙂

I was also really excited that, after more than a year of working completely remotely, we’ve found some ways to finally get together in person again. My personal favorites were when Mike Saparov, our VP of Engineering, took us sailing and when a group of us ran the California International Marathon together!

Milestone after milestone for Tecton & Feast

The data industry moves quickly and we’re growing with it! This year we’ve doubled down on building a reliable, simple enterprise Feature Store with Tecton, complemented by our open-source offering Feast.

• Tecton’s customer list grew by 5x! This growth was fueled by a new feature-development framework that brings together batch, streaming, and on-demand features. Over 90% of our customers use at least one set of streaming or on-demand features
• Feast was installed over 400,000 times, accelerated by the new lightweight Feast Feature Store that requires absolutely no backend
• Great teams such as Shopify, Twitter, Robinhood, and Salesforce have adopted Feast and are now valued partners contributing to it
• Tecton and Feast served over a trillion features to power ML predictions
• We shipped more than 150 new software releases
• Our apply() conferences brought together 2,500 ML practitioners – another testament to the hunger for knowledge in the world of Operational ML (complete with guided mediation and sing-alongs with Demetrios)
• We were recognized as a Gartner “Cool Vendor” 😎
• Our team appeared in more than 30 podcasts, meetups, and conferences.
• We became the first sponsor of the MLOps Community to show our support of the thriving ecosystem and watched it grow from 2,000 to 7,500 members
• Tecton now provides millisecond-fresh features on Kafka for real-time ML use cases such as fraud detection (blog). We’ve also published our first set of performance benchmarks that show how Tecton performs when serving >100,000 features per second

Looking forward to 2022

This has been a great year for us, and a great year for machine learning. But we’re still in the early innings of our journey, and we have big plans to continue making operational ML more accessible. With 2021 in the bag, what comes next?

We will continue to build partnerships across the ecosystem. Tecton will launch native support for the Modern Data Stack, starting with a deep Snowflake integration. We will also add support for Redis as an online store to achieve higher scale and lower latency predictions than is possible with DynamoDB. Through these partnerships, we will make Tecton seamless to use and more performant for our customers.

We also have big plans for Feast. We will continue to grow and invest in the open source community. And we will build a tighter relationship between Tecton and Feast, where both products will use the same framework to build ML features. Our team has seen thousands of models in production, and now we’re bringing that expertise into a single framework that makes it incredibly easy to use the most common patterns we see in ML features.

Finally, if 2021 was the year of the Feature Store, what will 2022 be the year of? Given the massive adoption of cloud-native data stacks and the maturing ML landscape, I expect 2022 to be the year of Operational ML – the year we’ll see exponential growth in the number of real-time ML applications that run in production to deliver real business impact and improve the end-user experience.

Our team will continue to grow significantly across the globe, and we’re hiring across many roles. If you want to build the future of ML with us and become part of the team, take a look at our careers page!

—
Kevin Stumpf
Co-founder and CTO