Hireable Β· Tashkent, Uzbekistan πŸ‡ΊπŸ‡Ώ

OnxxDatas

@Strong-Junior SWE  Β·  Writes code for food, vibe-coding 50/50
onxxdatas@laptop:~$
🐧 Ubuntu / Linux ✦ Vibe Coding w/ Claude πŸ•·οΈ Web Scraping 🐍 Python Backend πŸ›’οΈ PostgreSQL 🐳 Docker
01 // Identity

Who's behind
the terminal

I'm Abdulaziz - a Python backend developer from Tashkent, Uzbekistan who lives in the terminal and thinks in data pipelines. My world is split right down the middle: half Ubuntu/Linux power user, half AI-assisted vibe coder with Claude as my co-pilot.

I build backends that scrape the web, process data at scale, and serve clean APIs. Whether it's a Playwright async spider crawling thousands of pages or a FastAPI service powered by a Claude prompt chain - I ship things that work and work fast.

My coding philosophy: "Automate everything, own every line, and never leave the terminal unless absolutely necessary." Joined the WakaTime leaderboard Dec 2025 and the clock hasn't stopped since.

⏱ Live Coding Stats
WakaTime coding stats
View full profile β†—
πŸ‡ΊπŸ‡ΏTashkent, Uzbekistan Β· Since Dec 2025
3+
Years Python
50%
Vibe Coded w/ AI
∞
Terminals opened
200k+
Pages Scraped
02 // System

Living inside
Ubuntu & Linux

🐧
Ubuntu as Home Base
Ubuntu 24.04 LTS is my daily driver. Custom dotfiles, tiling window manager, terminal-first workflow. If it can be done in a shell, it's done in a shell. Zero GUI for dev tasks.
$uname -r
6.8.0-52-generic
$python3 --version
Python 3.12.3
$lsb_release -d
Ubuntu 24.04.1 LTS
βš™οΈ
Dotfiles & Dev Environment
Neovim + tmux + zsh + oh-my-zsh. Carefully tuned aliases, custom keymaps, and everything version-controlled. New machine setup runs via Ansible in under 10 minutes.
$cat ~/.zshrc | grep alias
alias py='python3'
alias dcu='docker compose up -d'
alias crawl='cd ~/spiders'
alias vibe='nvim .'
πŸ”§
Shell Scripting & Automation
Bash for everything repetitive: server health monitoring, log rotation, automated backups, cron-scheduled scrapers. Linux isn't just an OS - it's an automation platform.
$crontab -l
0 */6 * * * ~/scrapers/run.sh
0 2 * * * ~/backup/pg_dump.sh
*/5 * * * * ~/monitor/health.py
🐳
Docker-Native Everything
Every project containerised from day one. Docker Compose for local dev, multi-stage builds for lean production images. "Works on my machine" was never in my vocabulary.
$docker compose up -d
βœ“ postgres:16-alpine
βœ“ redis:7-alpine
βœ“ app:latest (FastAPI)
βœ“ celery worker
03 // Vibe Coding

50% code,
50% Claude

Vibe coding isn't lazy - it's leverage. I use Claude as a pair programmer, architect, and rubber duck all in one. Half my productivity comes from knowing exactly how to prompt an AI to get what I need, then reviewing and owning every line of output.

"Writes code for food, vibe-coding 50/50" β€” my WakaTime bio, and the most honest thing I've ever written.

My workflow: design the architecture myself β†’ prototype rapidly with Claude β†’ refactor and own every line. Result: ship 3Γ— faster without sacrificing quality.

I use Claude for: scraper boilerplate, async debugging, SQL optimisation, Dockerfile stages, API schema design, and code review. The AI writes the skeleton. I give it a soul.

claude_session.py Claude-assisted
# ✦ Vibe-coded with Claude, refined by hand

import anthropic
import asyncio
from dataclasses import dataclass, field

@dataclass
class VibeSession:
    client: anthropic.Anthropic
    history: list = field(default_factory=list)

    async def ask(self, prompt: str) -> str:
        """Shoot a vibe into the void, get code back."""
        self.history.append({
            "role": "user",
            "content": prompt
        })
        r = await self.client.messages.create(
            model="claude-opus-4-5",
            max_tokens=2048,
            messages=self.history
        )
        reply = r.content[0].text
        self.history.append({
            "role": "assistant",
            "content": reply
        })
        return reply

# Ship features 3Γ— faster. Own every line. βœ“
04 // Web Scraping

The art of
data extraction

Async Spider Engine Flagship
High-throughput async scraper with Playwright + asyncio. Handles JS-heavy SPAs, dynamic pagination, anti-bot bypass, and rotating proxies. 5k+ pages per hour on a single VPS.
Playwrightasyncioaiohttp2faRedis
Price Intelligence Bot
Automated competitor price monitoring across e-commerce platforms. Structured data stored in PostgreSQL, Telegram alerts on price drops. Runs in Docker on a cron schedule.
ScrapyPostgreSQLCUST-APIpython-telegram-bot
News Aggregator Pipeline
Scrapes 40+ pages, deduplicates with MinHash LSH, classifies topics, and uses Claude LLM to generate clean summaries. Exposed as a REST API via FastAPI.
RequestsFastAPIClaude LLMDeepTranslator
Data Enrichment Service
Takes raw business leads, scrapes public data sources, and uses Claude LLM to parse unstructured HTML into validated JSON schemas via Pydantic. Zero manual data entry.
PlaywrightClaude LLMPydanticFastAPIJSON
Typical Scraping Pipeline Architecture
🎯Target URL
β†’
🌐Playwright
β†’
🧩Parse HTML
β†’
✦Claude Parse
β†’
βœ…Pydantic
β†’
⚑API / Alert
β†’
πŸ—„οΈPostgres
05 // Arsenal

Full technical stack

🐍
Languages
Python 3.12SQLBashYAML
⚑
Frameworks
FastAPIPlaywrightCeleryPydanticSQLAlchemy
πŸ•·οΈ
Scraping
PlaywrightScrapyBeautifulSoup4httpxSeleniumaiohttp
πŸ—„οΈ
Data & Databases
PostgreSQLRedisMongoDBElasticsearchPandas
🐧
Linux / DevOps
Ubuntu 24.04DockerNginxGitHub Actions
✦
AI / Vibe Coding
Claude APIClaude LLMLangChainPrompt Eng.OpenClaw
06 // Contact

Let's build something
remarkable

I'm hireable and open to backend, scraping, or AI-integration gigs. Whether it's a spider, a pipeline, or a vibe-coded prototype β€” I ship it clean.