The alliance between Roomie and AWS has allowed us to solve the most complex problems of our installed base of clients in the commercial verticals of Data & AI and Robotics. The versatility of the different AWS tools has allowed Roomie’s expert software engineers to be able to implement our solutions in different commercial solutions.
We believe the future of Roomie and the world, lies in the implementation of robotics from organizations in order to optimize and automate their processes. For this reason we created Roomie’s Roomie Bot in 2018 with our own funding, investing in a R&D internal area to develop the first robots. However, we did not want to create a capital intensive business developing every single component and software tool from scratch. Instead, we focused on creating strong and strategic relationships with important organizations, such as AWS, to enable us to grow and scale at a much faster pace. This way, AWS has been an essential partner that has allowed our technology to be top of the line, while employing some of their state of the art, software modules.
At Roomie we have 3 robotics strategic verticals; Delivery Robots, Security Robots, and Social Robots. We utilize AWS RoboMaker in every robot we develop, as it represents a tool that already integrates other Amazon software programs (Amazon Kinesis, Amazon Rekognition, and Amazon Lex) and our robotics developers use ROS and ROS 2.
We adapt our robots with our four modular components internally developed at Roomie. Each module uses Amazon’s programs.
- Mobility Module (ROS 2, AWS RoboMaker): The patent pending technology we developed allows our robots to autonomously navigate through different terrains, avoiding obstacles and accidents. This opens up an infinity of use cases, as robots do not need an operator and robots take certain decisions to optimize productivity.
- Computer Vision (Amazon Rekognition): This module allows our robots, with the proper machine learning training, to identify objects, people, or even patterns.
- Legacy Integration (Amazon SageMaker): With this module, organizations can link their legacy systems with the robot, so it can be able to access the required information. We plan to develop AI apps that will allow the organizations to buy standalone already-developed apps.
- Voice User Interface (Alexa Skills, Amazon Lex): Essential for robots to communicate with end users. This module employs Amazon’s natural language processing so the robot can understand, process the information, and answer in a spoken manner, understandable by the end user.
Data & AI
Within our Data & AI solutions, we use a large number of tools developed by Amazon. At Roomie we offer six different Data & AI services.
- Machine Learning Solutions:
Using Amazon SageMaker we are able to model, build and train systems capable of solving problems with the most ambitious solutions in order to make your processes more efficient, or develop new projects from scratch.
We offer a very broad portfolio of solutions and with Amazon SageMaker we can accelerate innovation by implementing the latest technology and training algorithms for subtle implementation.
- Smart Assistants Omnichannel:
We develop intelligent artificial intelligence assistants available for different channels using two Amazon tools, Amazon SageMaker and Amazon Lex.
Amazon Lex allows you to enable a chatbot capable of interacting with the user via text or natural language.
- Data Lake and Advanced Analytics:
Enabling a Data Driven platform allows us to permeate the information from organizations to all partners to efficiently execute its activities.
In order to enable this service, we make use of AWS Lake Formation, which allows us to configure a secure data lake in a very agile period of time.
- Voice User Interfaces:
The voice user interface module enables communications between end users and computers employing Alexa Skills and Amazon Lex.
This way we can adapt robots or the organization’s conversational needs, in order for the user to be able to communicate directly with it.
- Computer Vision Solutions:
Amazon Rekognition is an excellent tool that we employ in order for robots, trained with machine learning and artificial intelligence, to detect from faces and people, to abnormal patterns or behaviours in a bank branch or a school.
We train the robots in order to recognize specific objects, patterns and images.
- Deep Learning and Reinforced Learning Solutions:
Employing a combination of Amazon Rekognition and Amazon SageMaker which allow us to improve predictions through various layers of artificial neural networks.
Our team implements reinforcement learning architectures and optimizes algorithms in order to maximize machine learning capabilities.