Case Study: Riskthinking.AI achieves scalable, auditable MLOps and faster, more reliable forecasting with Pachyderm

How Riskthinking.AI Uses Machine Learning to Bring Certainty to an Uncertain World

Riskthinking.AI, a forecasting and climate‑risk analytics firm, needed to move from ad hoc data science work to a production MLOps practice while building CovidWisdom to model pandemic policy scenarios for governments. Their team was spending too much time on manual data wrangling, versioning, and custom containers instead of modeling, so they selected Pachyderm as the platform to provide reproducible data pipelines and versioned data management.

Pachyderm delivered a data‑versioning and pipeline solution (accessed via pachctl) that let Riskthinking.AI run multiple models in parallel, visualize backtesting with full data lineage, and automatically promote the best‑performing model into the live application. The result was faster, more transparent experimentation, daily deployment of top models to CovidWisdom, clearer results for technical and nontechnical stakeholders, and a scalable foundation to process millions of data points for future climate‑risk forecasting — all enabled by Pachyderm.

RiskThinking