Let's talk
Let's talk

Our technology

We could have simply built a product, but we needed an ecosystem.

A male production techincian assembles pieces of the XperCount lid
XpertSea

IoT platform

Measurement technologies are the future of emerging industries, and we are proud to be spearheading that revolution in the aquaculture space.

At the core of our IoT platform is the Acquisition server, a daemon develop and optimized in C++. Interfacing high precision sensors through drivers and libraries, this server contains an embedded computer vision module that can run complex machine learning models. It exposes an API using websockets which can be used to control the device, locally or remotely.

Our devices are powered by Android and we leverage several out-of-the-box functionalities. This means we support most modern communications protocols and are able to push over-the-air updates.

IoT platform

Data platform

We're building the foundation of the first and largest aquaculture data source, a technological powerhouse that can be put to work to solve big problems.

NodeJS Services: API

api-nodejs-diagram.png

Python Services: Machine learning inference

api-diagram-1632487582.png

We designed our cloud platform to be efficient, robust, maintainable and reusable. Developed in Python & Node, it is 100% cloud-based, powered entirely by AWS, and built to scale.

Our APIs leverage EKS, providing fast response times while maintaining scalability.

Using AWS managed services, we built a skeleton for data pipelines. Those pipeline are designed to run compute intensive payloads like computer vision and deep learning model inferences.

SaaS platform

The SaaS platform is our customer-facing solution. A place to monitor daily information, and catch a glimpse of the future.

Our Software-as-a-Service platform is composed of every client, web app and tool that consumes data from the Data platform. This includes mobile ReactNative applications developed for Android/iOS and our React web application.

The Cloudfront Content Delivery Network serves up our web app, which retrieves its data from our serverless APIs on AWS.

SaaS platform

Machine learning

Machine learning is a powerful but complex beast. It can be overwhelming at times. We try to automate the boring parts and concentrate on building value.

ml-diagram.png

From classifying an aquatic animal to predicting the growth rate of a population, complex problems sometimes call for complex solutions. Hence, machine learning is at the core of our products.

We use different kinds of machine learning algorithms, from simple linear classifiers to recursive convolutional neural networks. All the model training and validation is done on the cloud.

We leverage AWS as an infinite source of power and use frameworks like OpenCV, Scikit Learn, Tensorflow and Keras. Our training framework can distribute and parallelize jobs on AWS EC2 clusters.

DevOps

We see DevOps as a process of continuous improvement. Every minute invested has a compounding impact. The more you put in, the more you get out.

Chart explaining how XpertSea uses DevOps

We started containerizing all our projects early on, utilizing Docker to abstract environment from execution. Most of us use Linux or MacOS, but we do have a Windows user (everyone has one of those.)

Everything we do is cloud-based. Whenever a developer pushes new code with git, a webhook triggers our CI. It builds the code, if needed, and runs the tests. If everything passes, it deploys the binaries to S3 or the new Docker to ECR. Otherwise, an alert is sent to the appropriate Slack channel.

Online services are deployed automatically by CloudFormation scripts which create all the AWS resources needed by the environment and fire up the machines.

All this, with a simple 'git push'.

Help us build the aquaculture intelligence platform of the future.

Line waves