Currently building · India
I am CRIXX
Builder. Systems thinker. Experimental technologist.
I learn by creating things that don't exist yet.
Builder. Systems thinker. Experimental technologist.
I learn by creating things that don't exist yet.
I'm a student. That's the official status. But the identity is this: I encounter something interesting and I can't stop until I understand it completely — and then rebuild it differently.
That compulsion pulled me into software, automation, and AI. I don't pick a subject and stay in it. I follow the question.
I build systems. I automate things that shouldn't need to be done twice. I experiment with AI not because it's trending but because it's genuinely one of the most interesting problems alive right now.
I'm early in this. Not pretending otherwise. But early isn't the same as unserious.
Jarvis is not a chatbot demo. It's a long-running personal AI system built around planning, execution, observation, memory, voice, local tools, and real desktop automation. The point is not to make it talk more — it is to make it understand what it can do, act safely, and get better at doing it.
A full-stack search engine with a provider-agnostic retrieval layer, normalized results, ranking, deduplication, multiple search verticals, history, API routes, and a lightweight frontend. Built to understand search as a system instead of hiding behind one provider.
A Windows telemetry, diagnostics, anomaly, and system-intelligence platform. It collects machine state, preserves history, exposes an API, and is being hardened around failure isolation, trustworthy signals, and a quiet local dashboard.
A manifest-driven .NET Windows toolkit for inspecting, configuring, repairing, and maintaining a PC with risk-aware execution, verification, checkpoints, desired-state tracking, and rollback. P1 → P20 is complete; real-machine acceptance testing is now finding the bugs unit tests can't.
A personal collection of tools and scripts I build to automate the tedious parts of my own life. Each one exists because I got annoyed enough to solve it.
Hands-on exploration of AI models, prompting strategies, and real applications. Not theoretical. Built, tested, honestly assessed.
Documented explorations of how things work under the surface. Software, hardware, business logic — anything complex enough to be worth understanding properly.
A backlog of standalone tools with one rule: they must be useful even if none of my other projects existed. Build the product first. Integrate later.
I'm interested in conversations about technology, systems, and what people are building. I don't want to network. I want to talk about something real.