Artificial intelligence and robotics are becoming increasingly integrated into our daily lives, and at the forefront of this revolution is Pyjam, a project founded by Pavel Malinovskiy. Pyjam is a unique platform that allows users to remotely control the actions of executors, generating data that is essential for training AI. What truly makes this project innovative is that every user, by controlling another person, essentially passes on their experience to shape future AI systems that will soon govern robots in our homes, on the streets, and in industries.
Pyjam: A New Level of Interaction
"We created Pyjam as a tool to eliminate geographical and physical barriers," says Pavel Malinovskiy. "But soon, I proposed to my development team that the interactions happening on our platform contained a colossal amount of data that could be used for training AI." In Pyjam, users can control the actions of an executor through a mobile application, giving real-time commands. This can be anything from a simple virtual tour to performing specific tasks.
"Every time someone controls an executor, they create something unique," Pavel continues. "It's not just about controlling movements—it's about sharing experience, which our AI analyzes and uses to make robots more adaptable and safer for all of us."
Hundreds of Thousands of Interactions — The Foundation for AI of the Future
Since its launch, Pyjam has recorded over a hundred thousand interactions between users and executors. These interactions serve as a foundation for AI training. "We use every element of these interactions—how the user controls the executor, what commands they give, how bystanders react. All of this helps AI systems become smarter and learn how to adapt to real-life situations," explains Pavel.
"The executors themselves are real people who also play a crucial role in this process. They respond to commands, interact with their environment, and our system records all these details. We are literally teaching robots how people behave in certain situations. This allows AI to learn from real human reactions and adapt in the future."
Pyjam: A Bridge to the Age of Robots
"When I founded Pyjam, I saw that interactions between people could become an excellent foundation for training AI," Pavel says. These interactions are already being used to train systems applied in robots such as Tesla Bot, Amazon Astro, Samsung Bot Handy, and others. "The robots of the future must not only follow commands but also understand how to properly react to unforeseen situations. And our job is to help them learn that."
"We are teaching AI how to predict human behavior, how to avoid collisions or discomfort in public spaces," Pavel explains. "For instance, if an executor approaches a group of people and they start reacting cautiously, our AI learns to identify such situations and adjust the behavior of robots. This is especially important when it comes to robots interacting with people in everyday conditions—whether it's delivering packages, helping in the home, or working in production."
National and Cultural Aspects: AI Learning Global Behavior
Pavel also highlights the importance of training AI to understand cultural and social differences. "One of the key points is that we can train AI to understand how people in different countries and cultures react to the same actions. This allows robots to adapt to local conditions," he says.
"In one culture, certain gestures may be perceived positively, while in another, they may provoke a negative reaction. Our data helps AI not only follow commands but also act appropriately, considering cultural and social norms. This speeds up the process of integrating robots into our homes and public spaces."
Learning AI Through Human Experience — A Path to a Safer Future
"When we talk about training AI based on data gathered through human interaction, it significantly accelerates technological development," Pavel notes. "Robots that can predict human reactions, avoid mistakes, and adapt to various conditions will become an integral part of our lives."
Pyjam is already helping to accelerate the arrival of the era of robots in every family. "My goal is to create a system that not only helps users remotely interact with one another but also trains AI. Thanks to our training data, robots will be truly useful and safe for people, which will soon change how we think about home and work," concludes Pavel Malinovskiy.
