Building on the successful generation of Orcasound open-source software through hackathons during 2018-2019, Orcasound applied in early 2020 to be a host organization for the Google Summer of Code (GSoC). Thankfully, we were accepted and joined an amazing array of open-source projects, many of which have benefited the bioacoustic community — like FFmpeg that generates our live audio streams.

We think we were selected not only because we had some track record as an cooperative open-source community, but also because we had an impressive, experienced team of mentors volunteer to guide our first students. Our team includes machine learning expertise from the University of Washington (UW) and Axiom Data Science, a 2019 GSOC student and mentors, bioacousticians studying Southern as well as Alaskan Resident Killer Whales, and volunteer coders from the Orcasound community:

• Valentina Staneva, Univ. Washington eScience (machine learning & data visualization)
• Shima Abadi, Univ. Washington Mechanical Engineering (acoustical oceanography & machine learning)
• Jesse Lopez, Axiom Data Science (computational data science & machine learning)
• Abhishek Singh, Final year Computer Science & Engineering student at NIT Durgapur, India/ GSoC’19 at ESIP.
• Hannah Myers, NGOS/Univ. of Alaska, Fairbanks (marine biology)
• Dan Olsen, NGOS (bioacoustics)
• Val Veirs, Beam Reach (Orcasound Lab hydrophone host, machine learning & noise analysis)
• Scott Veirs, Beam Reach (Orcasound coordinator, marine bioacoustics)
• Paul Cretu, Freelance software dev (lead Orcasound/orcasite dev for v1 UI) 
• Ivan Storck, Freelance full-stack developer
• Steve Hicks, Raspberry Pi expert and Orcasound orcanode pioneer

By March our team had put together a suite of project ideas that we thought would advance the Orcasound code base and/or contribute new, useful, open-source tools to the general bioacoustic and machine learning communities. Here are the topics we proposed, though we were also open to student-driven project ideas:

1) Active Listening & Learning of Orca Sounds

2) Real-time Orca Call Detection in No Time

3) Optimize Orcasound streaming software

4) Visualize human vs machine classifications in time & space

5) Visualize live sound within the Orcasound web app

6) Add and test icons in human detection UI within Orcasound web app

7) Add features to the Orcamap

In April we reviewed 34 (!) applications from undergraduate students around the world — from India and Pakistan to Europe and Russia, to Central and South America. It was a very impressive suite of proposals and we were hard-pressed to rank them, especially since — as a new organization — we were limited to selecting only two candidates. In the end — and with the Covid19 pandemic spreading uncertainty into everyone’s 2020 commitments, we chose Kunal Mehta and Diego Roderiguez both of whom had focused their proposals on the active learning project idea. Because their skill sets and goals were complimentary, and to hedge against the possibility that students and/or mentors might become less available than expected during the Covid era, we asked them to put together a joint proposal during the “community bonding” period in May.

We are very excited to be part of this vast educational effort that supports young coders around the world and advances the phenomena of open-source software development. And we can’t wait to see what the 2020 GSoC students build for Orcasound, citizen scientists, machine learning scientists, and bioacoustic researchers. You can follow their progress on the general Orcasound blog and/or the AI for Orcas project page #ai4orcas !