prodigy programmer · builder · independent research

I build the systems behind the tools.

I’m Reyaansh Sinha, a prodigy programmer building languages, virtual machines, AI infrastructure, operating systems, and tools that make software feel more understandable. Not just ideas. Real systems.

languages infrastructure AI systems
02 core public builds
AI infrastructure under test
Real systems over demos
Active builder mindset

My programming story started early. Then it kept going deeper.

My programming story started when I was around six years old, with Scratch. At first, it was just blocks, sprites, and small ideas. But to me, it felt like I had found a way to turn imagination into something real. I could make a character move, create a game, and build a world that actually responded to what I told it to do.

Around age seven, I started learning Python. That was when programming became more than just fun. I began to understand that code was not only for games. It could solve problems, automate things, and build real tools. From there, I kept moving deeper. I learned C, JavaScript, TypeScript, Rust, Go, and other languages because each one gave me a different kind of power.

Over time, I stopped wanting to only use other people’s tools. I wanted to understand how the tools themselves worked. That led me into building programming languages, virtual machines, operating systems, AI systems, and developer infrastructure.

One of my biggest projects was Splice, a programming language and virtual machine. Splice taught me how hard it is to build something from scratch, especially when bugs do not make sense at first. There were times when I spent days stuck on one issue, but every bug I fixed made me stronger. It showed me that real builders do not quit when something breaks. They learn why it broke.

Then I started working on AI infrastructure ideas like Halgorithem, AInfra, InfraVM, InfraOS, AgentVM, Plexus, and Moonshot Models. These projects came from one big feeling: AI and software tools should be reliable, understandable, and useful for real people. I did not want developers to suffer through confusing bugs, hallucinated answers, or tools that pretend to work but fail when they matter most.

That is also why I started Tangible Research Institute. My mission is to “Make AI Tangible.” I want to build the backbone: the systems, languages, verification tools, and infrastructure that other people can build on top of. Not everyone will see the backbone, just like not everyone thinks about the bottom blocks in Jenga. But when the backbone is weak, everything falls.

I am still young, but I do not see that as a weakness. I see it as proof that building starts before permission. I started with Scratch, then Python, then languages, operating systems, AI models, and research infrastructure. Every project has taught me something. Every failure has added another layer.

My programming story is not just about learning to code. It is about turning frustration into systems, ideas into tools, and imagination into something people can actually use.

Scratch at six Python at seven languages, VMs, OS work Make AI Tangible

Systems in the lab.

Splice

language + vm

Splice is a programming language and virtual machine built from scratch. It pushed me into the hard parts of parsing, execution, debugging, and designing something that feels small but has a real system underneath it.

The project taught me that broken systems are not dead ends. They are usually trying to explain exactly what you do not understand yet.

programming language virtual machine built from scratch

AInfra

AI infrastructure

AInfra is part of the Tangible Research infrastructure work: systems for making AI more reliable, understandable, and useful when the pressure is real. It sits near the backbone layer, the part most people do not see but everyone depends on.

It connects to the same mission as Halgorithem, InfraVM, InfraOS, AgentVM, Plexus, and Moonshot Models: make AI tangible, make tools explainable, and make builders less trapped by confusing failure.

github.com/TangibleResearch/AInfra AI backbone developer infrastructure

The part I keep coming back to.

AI is powerful, but not trustworthy by default.
Most people focus on making AI smarter. I focus on making it correct.
The future is not just generation. It is verification.

> models should explain themselves in ways a system can check

> confidence is not evidence

> correctness needs a standard, not a mood