A Journey to Robotics
Alexey’s career spans continents and industries, from co-founding an e-bike startup in Dubai to driving e-commerce growth in Singapore and London. Watching American Factory on Netflix sparked a realization about automation’s potential to transform repetitive work. “I felt genuinely inspired by the idea that robots could take over routine, uninspiring tasks, allowing people to focus on what truly requires creativity and intelligence. When I then got the opportunity to join a robotics project, it just clicked.” At Yango Tech Robotics, that click turned into commitment.
Being a Yangoer means you're building technology that matters, with people who care deeply about getting it right. There's no room for ego when you're teaching robots to work alongside humans.
What is Physical AI?
Yango Tech Robotics is advancing Physical AI, enabling robots to see, think, and act in real-world settings like warehouses. Yango’s Gen3 robots, guided by large models like LLMs, VLMs, or VLAMs, make autonomous decisions, sorting mixed items without fixed rules. “Physical AI is about giving machines not only perception and control, but understanding, allowing them to move from narrow, pre-programmed routines to adaptive, context-aware behavior: a step change comparable to what large language models did for software.” Yango's Vision-Language-Action (VLA) models integrate sight, language, and motion, letting robots understand simple instructions like "Pick up the blue box" or respond to gestures. A worker can say, "Move that to the shelf," and the robot grasps the intent instantly.
People can simply talk, gesture, or show what they want — and the robot understands the intent in context.
Yango Tech Robotics Now
From retail and e-commerce to pharmaceuticals, Yango robots manage thousands of products with over 95% accuracy, adapting quickly and fixing mistakes on their own.“Basically, robots can recognize and grasp thousands of SKUs, adapt to new items after just a few demonstrations, and recover from simple errors without human input.”The Specialist Picking Model is designed for real-world challenges:
- Versatile design using affordable RGB cameras
- Smart adaptation that adjusts strategies after failures
- Fast integration reducing setup from months to days
- Practical reliability for complex warehouse tasks
Opening the Specialist Picking Model
Yango Tech Robotics recently opened API access to its Specialist Picking Foundation Model, trained on over 2 million human demonstrations.
Openness drives adoption, and adoption drives improvement. Concretely, an integrator can now adapt our picking model to their custom gripper design in days rather than months.
This enables startups, integrators, and researchers to build on Yango's technology, fostering collaboration and speeding up industry-wide innovation.
Toward General-Purpose Robots
Yango's next ambition is a General-Purpose Robotic Foundation Model, that is versatile enough to switch from sorting packages to assembling products across different settings. “The hardest part isn’t scale — it’s diversity. To make a model that can reason and act across environments, you need data representing the real physical world in all its variety.”This pursuit embodies Alexey's favorite value: "Go Beyond":
Robotics is fundamentally about going beyond current limitations — beyond what's automated, beyond what's possible, beyond what anyone thought a machine could do.
Anti-burnout Hack
When asked about his anti-burnout hack, Alexey's answer is simple: rowing."It's repetitive enough to clear the mind yet demanding enough to reset the body. When you're synchronized with the motion and rhythm, the day's complexity fades—and clarity tends to return right after."From warehouses to everyday life, Yango's innovations are transforming how we work and interact with technology, making the future of robotics not just smarter, but more intuitive and collaborative.***We're rolling out our newsletter monthly, featuring interviews with our leaders from around the globe. Make sure to check our blog to stay in the loop!
