-
POSTECH
- Pohang || Seoul, South Korea
-
02:41
(UTC +09:00) - https://curling-gradient.dev
- @curling_grad
- in/ryang-sohn
- @curling-grad.bsky.social
- All languages
- Agda
- Arc
- Assembly
- Astro
- C
- C#
- C++
- CMake
- CSS
- Clojure
- CoffeeScript
- Common Lisp
- Coq
- Crystal
- Cuda
- Cython
- D
- Dart
- Dockerfile
- Elixir
- Emacs Lisp
- Go
- HTML
- Haskell
- Inno Setup
- Java
- JavaScript
- Julia
- Jupyter Notebook
- Kotlin
- LLVM
- Lean
- Lua
- MDX
- Makefile
- Markdown
- Meson
- Mojo
- Nunjucks
- OCaml
- Objective-C
- Objective-C++
- OpenEdge ABL
- OpenSCAD
- PHP
- Pascal
- Perl
- PostScript
- PowerShell
- Python
- Roff
- Ruby
- Rust
- SCSS
- Sage
- Sass
- Scala
- Shell
- Solidity
- Svelte
- Swift
- SystemVerilog
- TeX
- Tree-sitter Query
- TypeScript
- Typst
- VHDL
- Vala
- Verilog
- Vim Script
- Vue
- Zig
- mdsvex
- reStructuredText
Starred repositories
64-bit out-of-order superscalar RISC-V processor (RV64IMAFDV + Zba/Zbb) in SystemVerilog. Boots Linux 6.6.50 + BusyBox under Sv39 virtual memory and meets 500 MHz timing in Nangate 45nm.
[EMNLP 2026] PACE: Towards Surfacing Hidden Conflicts in User Requests
Generate Korean (Hangul) handwriting stroke-by-stroke with a GRU + Graves-style sliding attention window, learned from real pen recordings.
Gemma 4 12B baseline for RNGD — MOA 2026 competition baseline
🧩 Familiar patterns, beautiful Python types, better code. Zero runtime dependencies.
You are hanging out with Alice, Bob, and Charlie at Boolos' Brewery...
A high-performance inference engine for AI models
A curated resource list for learning AI performance engineering, from GPU fundamentals to production inference.
Rust compiler and standard library for FuriosaAI NPUs
Header-only C++20 parser combinator library, where everything is valid at compile-time
rootshell - The terminal, reimagined for Apple platforms
Unlocking _everything_ on the CPU with DRAM scrambling
Racing to the bottom of CPU performance
A very very very very very very very long interrupt
Fast and memory-efficient exact attention
FlashInfer: Kernel Library for LLM Serving
Ongoing research training transformer models at scale