Under the project's broadening-participation-in-computing (BPC) plan, A3EM introduces programming and conservation technology to students who might not otherwise encounter either, and shares acoustic monitoring know-how with practitioners internationally.
The curriculum builds on NetsBlox, a visual programming platform created at Vanderbilt University, and spans five half-day, project-oriented sessions. Each session introduces a conservation technology — such as wildlife cameras or GPS animal collars — followed by hands-on programming exercises. Instructors demonstrate each principle, but students complete individual tasks independently, from visualizing camera-trap photos from Mozambique's Gorongosa National Park to overlaying real animal-movement data on Google Maps.
Students at the June 2025 camp, Martin Luther King High School, Nashville
Average rated difficulty
Average rated usefulness
Average rated enjoyment
These ratings marked a significant improvement over the prior year's camp, attributed to the revised curriculum and a smaller class size. The curriculum has been published on the web, free for anyone to use.
Part of the team co-led a week-long workshop, "Bridging Knowledge Gaps to Combat Conservation Crime and Biodiversity Loss Using Acoustic Technologies in Thailand," bringing together government, academic, and non-profit participants to explore acoustic monitoring for local biodiversity solutions.
Participants from government, academia, and non-profits
Average rating for material and instruction quality
Signed between Colorado State University and Kasetsart University, Bangkok, to foster ongoing collaboration
Jesse Turner, the project's ecology PhD student at Colorado State University, has led prototype deployments across Alaska, Colorado, and Kenya — gaining hands-on training in instrumentation design, field logistics, and the collaborative relationships essential to long-term conservation research, alongside coursework including graduate-level machine learning.
Graduate students at Vanderbilt have advanced the project's embedded machine learning work, including quantized classifier design and adaptive filtering, gaining research experience that carries into their broader academic careers.
An undergraduate independent study explored environmental sound classification for microcontroller deployment, comparing state-of-the-art models (VGGish, YAMNet, OmniVec2) and fine-tuning a YAMNet-derived model to 91.25% accuracy at just 4 MB — small enough for onboard deployment.
Regular meetings between engineering, computer science, and field conservation specialists have generated new perspectives that are folded directly back into the education of students on both teams.