How Babylist Reclaimed 20% Engineering Time with Snowflake

Name: How Babylist Reclaimed 20% Engineering Time with Snowflake
Uploaded: 2026-10-04T11:57:46-04:00
Duration: 2 min 52 s
Description: Discover how Babylist reclaimed 10 to 20 percent of engineering time previously lost to tech debt by moving off MySQL and onto Snowflake's AI Data Cloud in partnership with Espresso AI. Espresso AI and Snowflake consolidated multiple Snowflake account...

Snowflake

10/04/2026

0 (0%)

Report Like Favorite

Transcript

to tackling tech debt. We now have that 10 to 20% of our time almost completely back. I'm Andy Byerly. I'm a senior analytics engineer at BabyList where we are helping growing families stay confident, connected, healthy, and happy as they kind of undergo this massive transition in their family's journey. My job on the data side is to make sure that the team's building these experiences for these families is backed by trusted data so that we know that we're doing the right thing to help these families grow the right way. Prior to ever using Snowflake, I was just using like a MySQL database to try and write very complex stored procedures to do analytics, which was miserable. We're able now to democratize our data in ways that we didn't think were possible a couple of years ago. And Espresso has been a huge unlock for us in reducing some of that overhead compute costs so that we can focus our engineering efforts away from optimization and more on doing what we do best, which is actually building the products and getting the data out to our stakeholders faster than ever. One big unlock that we have with Espresso specifically is the ability to cluster our warehouses together. So now we've been able to consolidate all of our Snowflake accounts into one instance, which has reduced a lot of the technical complexity underneath the covers and allows us again to focus on that time that we would normally spend managing all of those different environments and all those different variables. We can now just put right back into the product and really elevate our game. So our flagship product is DataDoola, which is a in-house AI data analyst that we've built on top of Snowflake's Cortex agents and semantic views. So what this does is it allows users to ask a question in natural language and get an answer back with consistent metric definitions and deterministic answers. What this unlocks for us is now we have all of our AI apps querying the semantic layer, so we get consistency across AI and BI. So no matter who you're asking or what you're asking, you're gonna get the same answer across all of these different platforms and across all these different tools. So this has been a really powerful unlock for our users because now we can go right into Slack and ping Cortex and get a real answer that we can action on right away. Snowflake will always be our semantic layer, source of truth, while Espresso's going to be kind of under the covers as our cost optimization layer, which means that as our AI usage scales, our build does not scale with it. We're able to spend that time focused on the families and not infrastructure. If I was joining a new data team who was evaluating different products, I would choose Snowflake every time.

TL;DR

Babylist reclaimed 10–20% of engineering time previously consumed by tech debt after migrating from MySQL to Snowflake's AI Data Cloud.
Espresso AI consolidated multiple Snowflake accounts into one instance, reducing infrastructure complexity and freeing engineers to focus on product development.
DataDoula, an in-house AI analyst built on Snowflake Cortex Agents and Semantic Views, delivers consistent, deterministic answers in natural language across all tools and platforms.
Espresso AI's cost optimization layer keeps compute costs controlled as AI usage scales, ensuring infrastructure spend does not grow proportionally with adoption.

Summary

In this short customer testimonial, Andy Byerly, Senior Analytics Engineer at Babylist, describes how migrating from a MySQL-based analytics environment to Snowflake's AI Data Cloud — in partnership with Espresso AI — transformed the team's productivity and data capabilities. Before the move, Babylist was spending 10 to 20 percent of engineering time managing tech debt and wrestling with complex stored procedures, leaving little bandwidth for product development. By consolidating multiple Snowflake accounts into a single instance using Espresso AI's warehouse clustering capabilities, the team dramatically reduced infrastructure overhead and reclaimed that lost engineering capacity. A centerpiece of the transformation is DataDoula, an in-house AI data analyst built on Snowflake Cortex Agents and Semantic Views. DataDoula allows business users to ask questions in natural language and receive consistent, deterministic answers grounded in a unified semantic layer — ensuring that whether a query comes through Slack, a BI tool, or an AI application, the answer is always the same. Espresso AI operates as a cost optimization layer beneath Snowflake, keeping compute costs flat even as AI usage grows. Byerly closes with an unambiguous endorsement: given the choice, he would choose Snowflake every time.

Chapters

0:00 - Engineering Time Reclaimed
0:38 - From MySQL to Snowflake
1:09 - Account Consolidation with Espresso AI
1:37 - DataDoula: AI Analyst in Natural Language
2:24 - Cost Optimization and Endorsement

Key Quotes

0:00 "We used to allocate 10 to 20% of our engineering time to tackling tech debt. We now have that 10 to 20% of our time almost completely back."
0:48 "We're able now to democratize our data in ways that we didn't think were possible a couple of years ago."
1:37 "Our flagship product is DataDoola, which is an in-house AI data analyst that we've built on top of Snowflake's Cortex agents and semantic views."
2:24 "Snowflake will always be our semantic layer, source of truth, while Espresso's going to be kind of under the covers as our cost optimization layer, which means that as our AI usage scales, our build does not scale with it."
2:44 "If I was joining a new data team who was evaluating different products, I would choose Snowflake every time."

FAQ

What is DataDoula and how does it work?

DataDoula is an in-house AI data analyst built by Babylist on top of Snowflake's Cortex Agents and Semantic Views. It allows users to ask questions in natural language and receive consistent, deterministic answers based on shared metric definitions — accessible directly from tools like Slack.

What role does Espresso AI play alongside Snowflake?

Espresso AI acts as a cost optimization layer underneath Snowflake, clustering warehouses together and consolidating multiple Snowflake accounts into a single instance. This reduces compute overhead so that as AI usage scales, infrastructure costs remain controlled.

Categories:

Tags:

Show more Show less

TL;DR

Summary

Chapters

Key Quotes

FAQ

Ensuring Compliance Through Audit Evidence: From CJIS to FERPA

Risk in Real Time Demo Series: Virtual Patching: Protection at the Speed of Exploitation

Harnessing Data Governance for AI with Cyera and Snowflake