// 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.
Custom Python weight loader project cover

Hands-On Build Guide: Custom GGUF & safetensors Weight Loader with Lazy Memory Mapping and Quantization in Pure Python

Learn how to implement a lightweight, lazy‑loading weight loader from scratch, complete with quantization support and memory‑efficient design.

November 15, 2026 · 10 min · 1985 words · martinuke0
Illustration of a PyTorch model with LoRA adapters

From-Scratch Parameter-Efficient Fine-Tuning Library: LoRA, QLoRA, and DoRA in Pure PyTorch

A hands-on guide to building a from-scratch PEFT library with LoRA, QLoRA, and DoRA in pure PyTorch, ready to showcase on your CV.

November 15, 2026 · 9 min · 1778 words · martinuke0
A tiled matrix multiplication visualization showing blocks of Q, K, and V matrices

Build a FlashAttention Kernel in Pure NumPy: Tiled Attention with Online Softmax

A hands-on guide to implementing FlashAttention in pure NumPy, demonstrating tiling, online softmax, and O(N) memory scaling — the exact systems skill hiring managers look for.

November 15, 2026 · 11 min · 2206 words · martinuke0
A laptop screen showing a Python script parsing a GGUF model file.

GGUF Parser and Dequantizer: A Hands‑On Portfolio Project

A step‑by‑step guide to parsing and dequantizing GGUF model files in pure Python, turning a niche tech skill into a tangible portfolio project.

November 14, 2026 · 9 min · 1916 words · martinuke0
Debezium CDC pipeline flowing from PostgreSQL to Kafka

Architecting Change Data Capture Pipelines with Debezium: From Postgres WAL to Kafka Topics

From PostgreSQL write-ahead logs to Kafka topics: how to build reliable, schema-aware CDC pipelines with Debezium.

November 14, 2026 · 10 min · 2031 words · martinuke0
Feedback