Data Science at Riskified

A glimpse into the changing world of human analytics at the onset of the AI revolution

Discover causal effects with retrospective data, address confounders, balance data using Propensity Score Matching, and access code

How can MLOps and continuous training bring real value to companies

Model replacement evaluation in a multi-model and dynamic environment

Explore some of the technologies, tools, and ideas we’ve been working on!

Learn how to leverage Airflow and Databricks to enable both teams to run independently while providing an end-to-end solution

A glimpse into problems solved using task forces

Four pitfalls to look for when using popular Machine Learning solutions on your data

Implementing a continuous training ML pipeline using MLOps ideas and shared tools

A glimpse into some of the technologies, tools, and ideas we have been exploring!

Why explainability (specifically using SHAP values) of a model is also important for the research phase

A key issue in many real-world problems: how can you know if your model is actually correct?

Metrics based on the Confusion Matrix shouldn’t be confusing. Here’s how I suggest to approach the KPIs definition challenge.

How do you get started with features in Data Science?