SoundTalks is a spin-off from KU Leuven university built on years of combined research in sound analysis and precision livestock farming. They created a hardware product for data capturing and combined it with a smart software layer to detect the spread of respiratory diseases in pig farms at an early stage.
We teamed up with the SoundTalks product and engineering team to build a high-load back-end infrastructure along with an accessible, holistic user experience that provides actionable insights to farmers and veterinarians to detect and prevent the spreading of diseases in farms sooner.
Hardware devices in pig farms collect raw sensor data, consolidate it via a local gateway and upload it onto the AWS cloud. An AI algorithm recognises coughs, sneezes, and other parameters to derive health insights that are reconvened to the farmers and veterinarians.
messages per day
git commits per year
automated unit tests run per commit
We’re active in a complex operational setting with limited internet connectivity and challenging digital adoption by our users. We found a great partner in Panenco given their end-user focus combined with strong engineering capabilities.
We fully analyzed the combined hardware and software flow for the SoundTalks ecosystem to be operational in farm settings.
After mapping user profiles and needs, we conceptualised:
(1) an onboarding flow with an installation procedure for the hardware devices and the related floor plan configuration
(2) the user interfaces for web and mobile apps where farmers and veterinarians can consult health insights and alerts
After clearing up the functional specifications we joined the SoundTalks team at the architectural drawing board with AWS at the center, getting ready for high data volumes and stringent data security requirements.
With this initial product definition in our back pockets, we were ready to define the product roadmap and get the software delivery teams up to speed.
Our teams opted for AWS infrastructure, currently adapting the setup to a serverless infrastructure with Lambda functions. The backend consists of REST APIs built in the Django framework using Python, with a React front-end web application and a React Native mobile app as the interfaces to serve the user base.
Efficient ci/cd pipelines with +90% test coverage both on the API and end-to-end make sure that we can release frequently and qualitatively. Last but not least we put performance checks in place to stress test all endpoints and monitor errors.
The SoundTalks product is being rolled out to all corners of the world. While the go-to-market and installations are taken care of by a commercial partner, we continue to support the technical operations and user experience. Besides helping the SoundTalks team grow in terms of software development maturity and the implementation of delivery culture, we assisted with the setup of user engagement instruments and a customer support system, while also guarding overall quality and user satisfaction.