Customers - LeadCrunch

The Challenge

LeadCrunch depends on regular updates to their ML models to deliver better recommendations to customers. But existing processes made updating models slow and ineffective.

Bi-annual implementation of ML models

LeadCrunch’s Data Science team regularly updates and refines ML models, but with 10+ sub-models and an ever-evolving stack, implementation was slow. Deployment of new models only occurred a few times a year.

Tough to get an overview of experiments

Each model was managed differently, which was inefficient. Prior to Verta, Data Scientists had difficulty gaining a holistic view of all of the experiments that were being run. This lack of insight further slowed the development of new models.

Maintaining models was a time-consuming process

Deploying and maintaining models often required large amounts of time and effort from LeadCrunch’s engineering team. But as a small company, these hours often weren’t available.

The Solution

Verta provides LeadCrunch with a unified MLOps platform that addresses their most pressing deployment and operations issues.

Faster implementation of new ML models

Verta inference allows LeadCrunch to easily deploy new models into production (batch, live, etc). “Previously, one of our main feature creation pipelines would take us 20 to 25 weeks,” says Alex Quintero, a Director of Data Science at LeadCrunch. “Once we brought Verta into the fold, that time cut in half. And it made it a lot easier to deploy and maintain our code.”

Accurate oversight of ML experiments

With model monitoring, experiment management, and a model registry, Verta’s MLOps platform supports the full model lifecycle. Verta’s intuitive experiment dashboard gave LeadCrunch’s Data Science teams confidence that they could get accurate, granular insights to the results of specific experiments.

Less time consumed by deployment and infrastructure

Verta freed LeadCrunch from the burden of worrying about deployment and infrastructure. “When you have a really small team like the one that we have, we can’t dedicate a lot of time and resources to the deployment of models,” says Brendan Homnick, LeadCrunch Sr. Data Scientist. “Verta takes all of that pain away.”

Verta + leadCrunch

LeadCrunch scales data science with Verta