Case Study: Digital Reasoning achieves scalable, reproducible, explainable ML pipelines with Pachyderm

Digital Reasoning - Customer Case Study

Digital Reasoning, a machine-learning and AI company serving clients from law enforcement to healthcare and finance, faced the challenge of processing large volumes of constantly changing, disparate data while maintaining explainability and reproducibility of models. To explore a next‑generation architecture with version‑controlled data and containerized pipelines, the team evaluated and adopted Pachyderm (running on Kubernetes) to bring code‑like versioning and pipeline management to their data workflows.

Using Pachyderm, Digital Reasoning’s team built scalable, repeatable, and explainable data‑science pipelines in a single day, automating data preprocessing, hyperparameter search, model tuning, model selection, inference testing, and reporting. Pachyderm enabled continuous end‑to‑end data ingestion with full provenance, integration with Jupyter notebooks for live experimentation, and rapid expansion from an initial pipeline to multiple use cases (including end‑to‑end audio transcription and image analysis), improving reproducibility, agility, and the speed of insight generation.

Digital Reasoning

Jimmy Whitaker

Manager