Case Study: CollectiveCrunch achieves 2–5× faster model development and more accurate wood inventory predictions with Valohai

Improving Smart-Forestry Through Machine Learning

CollectiveCrunch, maker of the Linda Forest SaaS for predicting wood quantity and quality, faced the challenge of turning large, diverse datasets (satellite, optical, LIDAR, process data) into reliable, monetizable ML models for the forestry supply chain. They needed to run extensive experiments, coordinate internal and freelance teams, and avoid loss of work from turnover. To manage this, CollectiveCrunch adopted Valohai’s ML orchestration platform for automated pipelines, version control, and parallel hyperparameter searches.

With Valohai, CollectiveCrunch separated workflows into pipelines (image cleaning, normalization, model training), ran parallel hyperparameter searches, and kept complete experiment history to enable collaboration and continuity. The result: Valohai accelerated model development by a factor of 2–5×, improved model selection and prediction accuracy, and reduced time spent on manual inspection—delivering more reliable, scalable insights to buyers and sellers in the forestry supply chain.

CollectiveCrunch

Rolf Schmitz

Chief Executive Officer