Security · Rust · Machine learning · Bioinformatics

Attacking assumptions, defending systems.

A developer who lives in the terminal: Python as the field tool, Rust as the craft, Linux as home ground. I study how systems fail — networks, models, living code — and build in the open so every claim here is checkable. The standing experiment is pulling machine learning toward Rust, where memory safety belongs to the language rather than a wrapper.

Open the console Get in touch Repositories
Self-avoiding walk on a lattice — the toy model of protein folding, and an ordinary graph traversal. Both readings are the point.
7 Public repositories
10 Certifications
9 Languages spoken
8 Locales on this page

01

Four disciplines, one toolset

Security · Rust · ML · Bio

None of these stands alone. Security is applied systems thinking, machine learning is code with statistics at the core, and biology is the open dataset that refuses to be simple. One toolset serves all four: Python, Rust, and a Linux that answers only to me.

Networks

Cybersecurity

Host discovery, port scanning and latency probes written as my own tooling rather than pasted from cheat sheets, backed by certified fundamentals in networking and information security. Linux is the environment I trust because its internals are readable.

Networking · InfoSec

Systems

Rust, as the craft

Ownership, traits and zero-copy design instead of wrappers. Python — three years of CLIs, pipelines and parsers — stays the field tool; Rust is where precision work happens.

Rust · Python

Learning

Machine learning, pulled toward Rust

Data work in pandas and NumPy, statistics as certified coursework, and a standing project: moving ML workloads into Rust so the performance and the memory safety come from the language itself.

pandas · NumPy · Rust

Reading

Computational biology, on the desk

Sequence alignment, structure and populations as a permanent side interest. Biopython for reading; the lattice at the top of this page is the toy model that keeps it interesting.

Bioinformatics

02

The record

Contact

Skills

Ten completed certifications

Data analytics, Python, version control, networking and security fundamentals — plus four Mimo tracks across JavaScript, Python, backend and frontend. Listed in full below, because a count without names is not evidence.

10

Voices

Nine languages, fluent where it counts

Persian, English and Azerbaijani at C2. German at C1, Turkish at B2, Arabic at B1 and still climbing, Chinese at HSK 2, French and Spanish at A1.

Persian - C2 English - C2 Azerbaijani - C2 Deutsch (German) - C1 Turkish - B2 Arabic - B1 Chinese - HSK 2 French - A1 Spanish - A1

9

01Google Data Analytics Professional2026
02Comprehensive Git & GitHub2026
03Python Programming, full track2024
04Computer Networking Fundamentals2025
05Information Security Essentials2025
06Git & GitHub Fundamentals2024
07JavaScriptMimo · 2026
08PythonMimo · 2026
09Frontend developmentMimo · 2026
10Backend developmentMimo · 2026
—Verification links pendingUntil each is linked, these are claims

03

What I can actually do

Claims sized to evidence

Code

RustPrimary craft · ML tooling

PythonField tool · 3 years

BashDaily driver

SQLJoins, aggregation

HTML / CSS / JSThis site, by hand

Machine learning & data

pandas / NumPyCleaning, aggregation

MatplotlibExploratory plots

StatisticsCertified coursework

BiopythonSequence reading

Security & systems

LinuxPrimary environment

TCP/IPCertified fundamentals

InfoSec essentialsCertified

PowerShellWindows administration

GitBranches, PR workflow

Nothing on this list says “expert”. Every label names the evidence behind it instead, because the fastest way to lose a technical reader is a claim they can disprove in one click.

04

Shipped

All repositories

Live · open source

This site

Shipped
Problem
Recruiters, collaborators and agencies all need the same facts in a different shape — kept in sync by hand.
Approach
One static site, no framework and no build step. A token-based design system, a hand-written interpreter for the console below, eight locales with right-to-left support, and a print stylesheet for clean A4 output.
Result
Down from 8.4 MB to about 70 KB gzipped, zero dependencies, WCAG AA contrast in both themes, and every drawing generated by a script in the repository rather than a model.
HTMLCSS Vanilla JSGitHub Pages
Source

Python · CLI

Terminal network toolkit

Local
Problem
Diagnosing a network meant running four utilities and correlating their output by eye.
Approach
A single Python command-line tool that runs host discovery, port scanning and latency probes concurrently on asyncio, then prints one report. Standard library only.
Result
Concurrency turns a serial subnet sweep into a parallel one. The speedup is not quoted here because it has not been measured under controlled conditions — the number goes up when the benchmark does, and not before.
Pythonasynciosockets
Publication pending

05

Console

Tab completes · ↑↓ recalls

Everything on this page is also reachable by typing. The interpreter is about two hundred lines of JavaScript with no dependencies, and it can reconfigure the site — try theme light, lang de or help.

ybreakless@portfolio type help

06

Get in touch

Same handle everywhere

Want to know who I am? Message me on social media — the handle is @ybreakless everywhere. If you need a name to call me, ask: I go by Ario.

GitHub@ybreakless LeetCode@ybreakless Hugging Face@ybreakless