G2

Steven L.

Cloud Security Architect

Financial Services

Mid-Market (51-1000 emp.)

9/1/2026

"Snorkel’s Programmatic Labeling Transformed Our Threat Detection Workflow"

5/5

What do you like best about Snorkel Predictive ML?

I've been working with Snorkel Predictive ML for about six months now, mainly for improving our threat detection models at FinSecure Bank. The programmatic labeling approach is genuinely different from anything else I've tried. Instead of spending weeks manually tagging thousands of security logs and network traffic patterns, I can write labeling functions that encode my domain expertise directly into the training pipeline. Review collected by and hosted on G2.com.

What do you dislike about Snorkel Predictive ML?

There's definitely a learning curve here, and I'm saying that as someone with a strong ML background. The weak supervision paradigm takes time to internalize, and creating effective labeling functions that don't introduce bias requires careful thought. Some of our security analysts found the initial setup intimidating. Review collected by and hosted on G2.com.

What problems is Snorkel Predictive ML solving and how is that benefiting you?

The biggest problem Snorkel addresses for us is the bottleneck of labeled training data for security models. In financial services, we have vast amounts of log data, network traffic, and transaction records, but creating labeled datasets for threat detection is incredibly time-consuming and expensive. Manual labeling requires security experts who are already in short supply, and the labeling quality can be inconsistent. Review collected by and hosted on G2.com.