About the Role
Kashmir World Foundation is a 100% volunteer-driven nonprofit using AI and bioacoustics to protect endangered species and ecosystems. Our flagship project leverages sentinel species as biological sensors, analyzing animal vocalizations and soundscapes to detect environmental and human-driven threats before they escalate to poaching or wildlife loss.
We apply machine learning to high-dimensional acoustic embeddings, modeling temporal dynamics, regime changes, and anomalies to enable proactive conservation.
🔗 Learn more:
Sentinel Species as Guardians: https://www.kashmirworldfoundation.org/post/sentinel-species-as-guardians-a-breakthrough-in-bioacoustic-conservation
Documentary Beyond the Noise: https://youtu.be/-Ti0vHeNF8o
Internship Overview
We are seeking an NLP / ML intern to analyze bioacoustic embeddings from field data (Costa Rica expedition), focusing on temporal structure, stability vs. transitions, and early threat detection.
What You’ll Work On
Analyze high-dimensional audio embeddings
Apply PCA, UMAP, and trajectory analysis
Study embedding geometry and regime shifts
Build reproducible ML experiments
Explore models for behavioral and environmental change detection
Requirements
Required Skills
Dimensionality reduction (PCA, UMAP)
Understanding of embedding spaces and temporal dynamics
Experience with reproducible ML workflows
Familiarity with HMMs, change-point detection, or regime-switching models
PyTorch and/or TensorFlow
Nice to Have: Self-supervised or contrastive learning, edge deployment, conservation tech interest