The lives and work of software engineers and conservationists might seem worlds apart. However, last month the two came together at the Zoohackathon, hosted by ZSL London Zoo, in collaboration with Bloomberg and the U.S. Embassy London, to develop technological tools to protect endangered species around the world.
The London Zoohackathon is part of a larger global initiative held at zoos across the globe that is aimed at finding innovative tech-driven solutions to help tackle illegal wildlife trade. Now in its second year, the initiative is organized by the U.S. Department of State and the Association of Zoos and Aquariums (AZA).
In addition to sponsoring this year’s Zoohackathon, Bloomberg invited nine of its London-based engineers to volunteer their time and talent and participate on different teams in the event. Through the company’s philanthropy and engagement program, Bloomberg employees are offered a range of skills-based volunteering opportunities to enable them to apply their expertise to support non-profit organizations in contexts outside their day-to-day responsibilities.
With more than 5,000 engineers and data scientists globally, and 800 in London alone, support and participation in social hackathons is a key strand of the company’s philanthropy and engagement program. By harnessing these talented engineers’ skills and resources, the company is able to help benefit society and also expand its employees’ ongoing professional development.
Bernát Gábor, a software engineer in Bloomberg’s London office, participated in both this and last year’s Zoohackathon. He worked alongside team members with similar backgrounds in coding, but also with participants specializing in human resources, sales and marketing. Last year, his team won the London event by prototyping a content management system that could alert travelers to items they might be asked to buy that are the result of illegal poaching.
Gábor says non-technical participants play a big role in helping ensure the Zoohackathon’s projects are geared towards the real world. The non-coders on his team, says Gábor, talked to zoo visitors about the potential product, while the coders were trying to build it. “We were able to know how viable something truly would be, instead of just trying to create something and then needing to sell it,” he explains.
Since many of the hackers come to the zoo with little knowledge of endangered species, the first evening is spent with speakers and subject matter experts who provide the participants with a broad overview of the challenges faced to get them up to speed as quickly as possible. Those same experts are available throughout the weekend to provide advice and ensure that the proposed solutions address the problem in a practical and useful way. That means the projects have to be inexpensive, without significant energy cost, and easily deployable and reliable in remote environments. “It’s really important to have people there who can talk about those requirements,” says Rachael Kemp, one of the hackathon’s organizers and a conservation technology project manager at ZSL.
This year, Gábor chose to work on a project that would help rangers better manage potential threats from the illegal wildlife trade. “Whenever an incident occurs, the proposed system would call one of the rangers, and the ranger could accept the mission and follow up to investigate a potential poaching threat,” says Gábor. His team used public APIs for sending voice messages, and built an app server that could run on a Raspberry Pi computer. The system makes calls automatically, and creates a log. “One of the big problems in the field is corruption,” says Gábor. “A logging system can help you follow up with everything.”
The winning team of this year’s London event – ODINN (Onsite Dynamic Identification Neural Network) – of which Bloomberg engineer Krishna Raj Sapkota was a member, aims to improve image recognition so that software can better identify poaching risks. As it stands, rangers get slews of incoming images that may–or may not–represent threats to wildlife. “Someone has to sift through those,” says Kemp. An image-recognition system that could flag the most critical risks would help rangers address those issues in real-time.
“The simplicity of the solution lies in the fact that it requires no additional hardware and can be retrofitted to camera traps already deployed,” adds Sapkota.
The hackathon gives zoo staff access to ideas that will have real impact in the field, says Kemp. But it also connects the zoo with people with whom it doesn’t regularly work. “It’s a great opportunity to get these people in, and get them excited about what we’re doing,” says Kemp. Or, as Gábor put it, “The whole event just has a really good vibe.”
Bloomberg has proudly supported ZSL since 2009 and has supported the Zoohackathon from its beginning. Bloomberg’s commitment to using data science and human capital to solve problems at the core of society is part of a long tradition of advocacy for using data.