// 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 sleek notebook with code and diagrams

From‑Scratch MoE Transformer Block with Sinkhorn Routing in NumPy

A hands‑on guide to building a Mixture‑of‑Experts transformer block from scratch in NumPy. It features Sinkhorn‑based routing and balanced expert load, ready to showcase on a CV.

November 22, 2026 · 8 min · 1554 words · martinuke0
Mixture-of-Experts router architecture diagram

A Hands-On Build Guide: Mixture-of-Experts Router with Top-K Gating and Auxiliary Load-Balancing Loss

A complete guide to building a Mixture-of-Experts router from scratch, including top-k gating, auxiliary load-balancing loss, runnable Python/NumPy code, and production-ready extensions for your CV.

November 22, 2026 · 11 min · 2171 words · martinuke0
Diagram of a mini GPT trainer with autograd and all-reduce

Build a Mini GPT Trainer: Hand‑Roll Autograd + Data‑Parallel All‑Reduce in Pure NumPy

Learn how to implement a tiny GPT trainer in pure NumPy, complete with autograd and distributed gradient all‑reduce, and why it signals real systems skill to hiring managers.

November 21, 2026 · 9 min · 1806 words · martinuke0
Cover image for paged KV-cache project

From-Scratch Paged KV-Cache with Block-Table Virtual Memory and Eviction Policies in Python

A hands‑on guide to building a paged KV‑cache in Python, complete with virtual memory, eviction policies, and benchmark results you can run today.

November 21, 2026 · 8 min · 1662 words · martinuke0
SPIFFE trust diagram securing inter-service communication

Implementing SPIFFE Workload Identity for Zero Trust Networks

SPIFFE replaces static credentials with dynamic mTLS for workloads. This article covers deployment patterns, service mesh integration, and hardening strategies for zero trust networks.

November 21, 2026 · 7 min · 1380 words · martinuke0
Feedback