How PellencST Cut Time-to-Model by 90%
From 6-9 months to 1 month: PellencST gained a competitive edge in the sorting machine market with Picsellia.
Get Similar ResultsPellencST is a pioneering waste sorting machine manufacturer operating over 2,500 machines across 40+ countries. They collaborate with major environmental organizations like Paprec, Veolia, and Suez, serving as the market leader in France and a key international player in recycling technology.
PellencST faced growing demand for complex sorting capabilities that traditional sensors couldn't handle. They needed to internalize AI development to stay competitive.
Growing demand for complex capabilities like silicone cartridge sorting and improved paper/cardboard separation that regular sensors cannot handle.
Initially relied on external CV model development, creating bottlenecks when clients requested customizations or new capabilities.
Model development took 6-9 months from start to finish, with annotation alone requiring a full month.
Sorting machines require real-time AI processing. Needed to internalize capabilities while maintaining performance.
From challenges to solutions
- —6-9 months for model development
- —1 month for annotation alone
- —1-2 months for data acquisition
- —Relied on external CV development
- —Bottlenecks in customization requests
- 1 month total time-to-model
- 2 weeks for annotation
- 3-4 days for data acquisition
- Full internal AI capabilities
- Rapid response to client needs
“Using Picsellia has changed how we develop, monitor, and deploy our models. Complex CV applications are more reliable and reach production faster.”
How PellencST uses Picsellia
Pre-annotation and structured campaigns cut labeling time in half
All AI tools consolidated in one platform for rapid iteration
How Picsellia delivered
Picsellia unified PellencST's MLOps workflow, enabling them to internalize AI capabilities and respond rapidly to market demands.
Unified MLOps Platform
AI Laboratory →All AI tools consolidated in one location. Centralized repository for models, datasets, and experiments.
Pre-Annotation Capabilities
Labeling Tool →Accelerated data preparation with AI-assisted labeling. Annotation time cut from 1 month to 2 weeks.
Model Lifecycle Management
Model Monitoring →Complete monitoring and management of models across 2,500+ machines worldwide.
Open-Source Integration
Automated Pipelines →Seamless integration with existing tools like Airflow and MLFlow for custom workflows.
Business impact
PellencST transformed their competitive position through dramatically faster AI development and deployment.
Time-to-model reduced from 6-9 months to just 1 month for training, testing, and validation.
Faster customized solution development enabled PellencST to win major contracts.
Technological innovation through AI positioned them ahead of competitors.
Complete internal capability for model retraining and deployment without external dependencies.
Ready to achieve similar results?
See how Picsellia can help your team build and deploy computer vision models faster.