Conservation needs better technology. Most organizations can’t build it themselves.
We build it for them. Open-source machine learning tools for counting fish, detecting fires, identifying animals, and monitoring ecosystems. Free to use. Free to modify. Built with the people who use them.
Better tools. Better data. Better outcomes.
🛠️ What We Offer
We build custom conservation tools adapted to your specific needs and datasets, empowering organizations with actionable environmental insights. Our expertise spans the full development lifecycle—from data collection and model training to deployment and maintenance.
Custom AI/ML Model Development
We design and train machine learning models tailored to your conservation challenges, whether it's species identification, behavior detection, or environmental monitoring.
Data Pipeline & Processing
We build efficient systems to process large-scale datasets — terabytes of audio, video, or sensor data — turning raw information into meaningful insights.
Deployment & Integration
We deliver production-ready solutions that integrate with your existing infrastructure, from edge devices in the field to cloud-based platforms.
Open-Source Tools
Most of our work is open-source, ensuring transparency, reproducibility, and long-term sustainability for your organization.
🔍 Concrete Examples
Tracking forest elephants
Enables monitoring of populations and detection of poaching threats
Developed an acoustic analyzer that processes years of audio recordings at scale, automatically detecting elephant rumbles and gunshots in continuous soundscapes.
Monitoring salmon migration
Enables precise population tracking
Built SalmonVision — a computer vision system using underwater cameras, sonar, and drones to automatically count and classify salmon species in real-time across multiple river sites.
Detecting forest fires early
Detecting fires from 35km away, protecting forests in Europe
Enhanced Pyronear's ML model for real-time fire detection from tower-mounted cameras, minimizing false alarms while maintaining high sensitivity to early smoke signatures.
Identifying individual bears
Enables population monitoring, and behavioral research
Built a non-invasive facial recognition system using instance segmentation and metric learning to identify individual bears from camera trap photos without physical tagging.
Explore more of our work on our projects page.
Have a conservation challenge? Get in touch, we'd love to explore how we can help.
đź‘‹ Meet the Team
Arthur Caillau
Building open-source AI tools that empower conservationists worldwide. Passionate about making cutting-edge technology accessible to protect biodiversity and drive real-world conservation impact.
Jeremy Vuillermet
Crafting intuitive interfaces that put powerful conservation technology into the hands of field researchers and wildlife experts. Focused on creating seamless user experiences for complex AI-driven tools.
Eelke Folmer
Bridging ecology and technology through data science, geospatial analytics, and machine learning. Leverages drone imagery and photogrammetry to deliver actionable insights for conservation projects worldwide.
Thor Veen
Evolutionary biologist turned conservation technologist. Combines expertise in biodiversity monitoring with cutting-edge computer vision and machine learning to develop innovative wildlife tracking solutions.