About — Vedant Neve

👋 Who am I?

I'm Vedant — I want to learn everything about everything. I build things that make me go "Huh, that's interesting!" and "Wow, I didn't know that was possible!", and then I write about them here. I have a soft spot for compilers, low-level programming, GPU graphics, and performance optimization — and I'm always up for exploring new domains and technologies.

My primary focus is on compiler engineering — I'm an active LLVM contributor, working on things like GVN optimizations, SelectionDAG pattern matchers for funnel shifts, and libc refactors. My research interests revolve around compiler optimizations, program transformations, static analysis, LLVM pass infrastructure, parallelizing compilers, and IR-level code generation. I'm currently exploring source-level CUDA↔ROCm transpilation for portable GPU compute.

On the projects side, I've built Chakravyuha — an LLVM-based code obfuscation framework with control flow flattening, polymorphic string encryption, and dead code injection; obvcc — a multi-language C compiler (Rust lexer + OCaml parser → x86-64 assembly); Raylib-CUDA — a zero-copy GPU compute interop library using OpenGL DMA; and obv-OS — an operating system from scratch. I also dabble in HTTP servers in pure C, terminal media players in Go, and WebAssembly experiments.

I lead the Competitive Coding Club at my college and co-lead the DSA & CP track for Google Developers Group on campus — organizing ICPC-style contests and leading technical sessions on algorithms and data structures.

⚡ Tech Stack