The Kid with a Laptop
It started in Punjab, 2017. No Silicon Valley garage, just a 12-year-old kid who didn’t even have a phone, deciding that if he couldn't have the tech, he’d build it. From those first lines of code to a life-changing Oracle GPU Grant, the mission was always simple: Zero excuses, pure innovation.
Personal Over Powerful
While billion-dollar labs fought for benchmarks, Vansh fought for connection. He realized a static LLM could never be a true friend because it forgets. A real companion remembers shared experiences. This obsession with personalization led him to a breakthrough: Inference-time weight updates.
The Architecture of AGI
In 2025, that search for a better companion uncovered a new paradigm: Continual Learning. By applying scaling laws to this architecture, we didn't just build a personalized model—we built the framework for a Human Brain in code. A model that doesn't just process; it learns, adapts, and innovates in real-time.
The Global Frontier
Today, we are a Country of Geniuses competing with the world’s biggest labs from a single room in India. We are building a future of decentralized, self-improving devices in every pocket.
Innovation has no borders. Greatness needs no permission.
Successes That Define Our Path to AGI
Vision CLv1
Achieved the "Last Mile" of AGI; the birth of Continual Learning, allowing the model to learn and innovate in real-time.
2026
The Soul of Speech
Built native Speech-to-Speech (S2S) capabilities; launched Rose, achieving SOTA results in emotional texture and latency.
2025
Multimodal Birth
Added vision and audio perception; Vision evolved into a native multimodal system that "sees" and "hears".
2024
Vision LLM v1
Trained the initial version of Vision; outperformed GPT and Gemini in core reasoning benchmarks.
2023
The Foundation
Early experimentation with transformers; training the first foundational model architecture using the initial research grant.
2022
