Skip to content
View RPegoud's full-sized avatar
22 followers · 6 following

Achievements

Achievement: StarstruckAchievement: Pull Sharkx2Achievement: YOLO

Achievements

Achievement: StarstruckAchievement: Pull Sharkx2Achievement: YOLO

Block or report RPegoud

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.

Content in all repositories owned by your account will be closed.
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
RPegoud/README.md

Hi, I'm Ryan 👋

ML research engineer working across reinforcement learning, post-training, and hardware-aware ML. MSc in Computational Statistics and Machine Learning from UCL, previously a visiting researcher at BMW on end-to-end autonomous driving.

⚙️ Currently building: kompass, a harness for writing, testing and benchmarking CUDA and Triton kernels against PyTorch baselines (online softmax, fused linear cross-entropy, FlashAttention, GEMM next). Release coming soon.

🧠 Interests: Open-endedness, RL and post-training, automated research, inference efficiency and GPU kernels, vision-language-action models

✍️ I write about RL, JAX and GPU programming on Towards Data Science

📫 ryan.pegoud@outlook.com · Portfolio · LinkedIn

Tools

JAX · PyTorch · Triton · CUDA · vLLM · Modal · Polars · Hugging Face

Pinned Loading

  1. kompass kompass Public

    Python tooling suite for GPU kernel development

    Python

  2. jym jym Public

    JAX implementation of RL algorithms and vectorized environments

    Jupyter Notebook 50 2

  3. Hercules Hercules Public

    [UCL MSc Thesis] Neural Memory Augmentation of Large Language Models

    Python