Skip to content
View Peterc3-dev's full-sized avatar
1 follower · 1 following

Block or report Peterc3-dev

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Peterc3-dev/README.md 
 ██████╗ ███████╗████████╗███████╗██████╗  ██████╗██████╗
 ██╔══██╗██╔════╝╚══██╔══╝██╔════╝██╔══██╗██╔════╝╚════██╗
 ██████╔╝█████╗     ██║   █████╗  ██████╔╝██║      █████╔╝
 ██╔═══╝ ██╔══╝     ██║   ██╔══╝  ██╔══██╗██║      ╚═══██╗
 ██║     ███████╗   ██║   ███████╗██║  ██║╚██████╗██████╔╝
 ╚═╝     ╚══════╝   ╚═╝   ╚══════╝╚═╝  ╚═╝ ╚═════╝╚═════╝

Building open-source ML tooling for AMD APUs. Squeezing every FLOP out of hardware the industry forgot about.

I found 3 bugs blocking ML on AMD's newest GPU architecture (gfx1150/RDNA 3.5), built a public PyTorch wheel for it, and root-caused an NPU driver bug on Strix Point (amd/xdna-driver #1257).

Currently prototyping a routing system for CPU + GPU inference on AMD APUs (+ NPU when runtime support lands), using hyperdimensional computing. HDC routing showed 21% throughput improvement in Phase 1 benchmarks.

Rust · Python · Kotlin · JavaScript · C++ · Shell CachyOS · Docker · Chrome Extension API · Vulkan

CEH in Training · HackerOne Researcher · Building AgentSpyBoo Phase 3

Bugs I Found

Things I Shipped

Former ATP Training Partner (Federer/Djokovic). Yes, really.

Pinned Loading

  1. agentspyboo agentspyboo Public

    First Rust-based AI red team agent on AMD XDNA 2 NPU — portable, air-gapped, zero-cloud autonomous pentesting

    Rust 1

  2. d-board d-board Public

    Ottholinear Android keyboard - true grid layout for improved typing ergonomics.

    Kotlin

  3. retune432-android retune432-android Public

    Android app to batch-convert audio from A440 to A432 Hz with metadata preservation

    Kotlin

  4. torch-vulkan torch-vulkan Public

    Vulkan compute backend for PyTorch — runs on any GPU. PrivateUse1 dispatch, SPIR-V shaders, zero ROCm/CUDA dependency.

    C++

  5. omnivoice-gfx1150 omnivoice-gfx1150 Public

    Running k2-fsa/OmniVoice voice cloning TTS on AMD Radeon 890M (gfx1150, Strix Point) integrated GPU — report, benchmark, and reproduction guide

    Python 1

  6. recursive-routing-racer-rs recursive-routing-racer-rs Public

    Rust Vulkan inference runtime — 15.8 tok/s on Radeon 890M. Zero-ceremony dispatch, batched command buffers, Q4 quantized matmul.

    Rust