// TODO: I’m martinuke0

Welcome to my corner of the internet. This website is a personal blog which I use as a platform to document my learning journey and showcase it for the world to see.
A visual representation of tiled attention computation

FlashAttention Kernel from Scratch in Pure Python with Tiling and Online Softmax

This tutorial walks you through implementing a FlashAttention kernel from scratch, using tiling and online softmax to achieve memory-efficient attention. You’ll gain hands-on experience with low-level systems techniques that hiring managers value.

November 19, 2026 · 6 min · 1111 words · martinuke0
LoRA adapter fine‑tuning on a sample image dataset

Hands-on LoRA Trainer from Scratch in Pure Python

A hands‑on guide to training LoRA adapters with pure Python, no heavy frameworks, ready for CV portfolio projects.

November 19, 2026 · 8 min · 1634 words · martinuke0
A distributed system architecture diagram showing CI workers connecting to a remote cache cluster

Architecting Hermetic Bazel Remote Caching for Reproducible CI Builds at Scale

Bazel’s remote caching can turn your CI from a slow, flaky bottleneck into a fast, deterministic feedback loop, but only if you architect it with hermeticity and scale in mind.

November 18, 2026 · 8 min · 1569 words · martinuke0
Python code representing a byte-level BPE tokenizer

Byte‑Level BPE Tokenizer from Scratch in Pure Python

A hands‑on guide to implementing a GPT‑2‑style tokenizer from scratch, complete with runnable Python code and round‑trip byte decoding.

November 18, 2026 · 10 min · 2020 words · martinuke0

Pure‑Python GGUF & Safetensors Weight Loader

A hands‑on guide to writing a pure‑Python GGUF/Safetensors weight loader that reads binary headers and returns NumPy arrays, perfect for demonstrating low‑level ML engineering skill on a CV.

November 18, 2026 · 8 min · 1529 words · martinuke0
Feedback