Python Backend Developer • Integrations & Automation

W1bbex Python Backend Developer Automating manual workflows

I connect services, build bots and create tools for automation.

developer_profile.py

projects = {
      "iot":        "PC Control",
      "ast_tool":   "TG-FlowMap",
      "mobile_iot": "Phone Control",
      "devops":     "VDS Control",
      "ai_cli":     "W1x CLI",
      "fpc_core":   "FPC-AutoOffline",
      "fpc_smm":    "FPC-AutoSMM",
      "fpc_test":   "FPC-E2E",
      "data":       "FunPay Parser",
      "platform":   "W1bbex Hub",
      "webgl":      "W1bbex AI Universe",
  }

01 / About

About me

I’m a 16-year-old student software developer. I have been writing code since January 2026, progressing from Python basics to production systems, PyPI packages, ESP8266 hardware control and Cython builds.

I focus on backend logic, integrations and automation, using Telegram Bot API, external APIs, Linux, systemd, SQLite, ETL and AI assistants.

I build services for order fulfilment, plugin distribution, isolated E2E testing and remote server control, turning ideas into working products.

02 / Projects

Projects and tools

Open each card for architecture and implementation details.

Technical detailsClose details

Problem

Remote desktop software requires an active session; keyboard dongles introduce security concerns.

Solution & process

An ESP8266 controls an ATX optorelay and reads Morse input. A Windows daemon uses PsExec and Win32 APIs to work with the secure Winlogon desktop without USB keyboard emulation.

Scale / metrics

Relay response below 1.5 seconds; 412 automated tests (all passed locally); VPN node rotation; Yandex Music control; C# sensor collector; over 117,000 SQLite WAL telemetry points.

Architecture Telegram UI ⇄ Cloud VDS (TLS) ⇄ Mosquitto MQTT (QoS 1) ⇄ [ESP8266 (OOB) + Winlogon Session 0 + Desktop Agent]

Lessons & development

Session 0 isolation, WTS session tracking, interrupt handling, EEPROM brute-force protection and extraction of Telegram custom emoji icons.

Technical detailsClose details

Problem

Growing bots have complex routers, callbacks and inline keyboards that are hard to audit.

Solution & process

A Python AST analyser maps screens without executing code, while a Telethon crawler explores live bots using Dijkstra’s algorithm and filters payment buttons.

Scale / metrics

Approximately 50 ms static analysis, 217 automated tests, Telegram API limit validation and an interactive DOM/SVG graph.

Architecture [Python AST (SAST) + Telethon Dijkstra (DAST)] → Unified BotMap Spec → GPU DOM/SVG Graph & AI Benchmark

Lessons & development

Resolves dynamic f-strings, conditional expressions and defaults; prevents graph collisions and optimises text rendering.

Technical detailsClose details

Problem

Alarms and lecture recording depend on manual work while Android restricts background services.

Solution & process

A Termux daemon schedules alarms, records AAC/M4A audio, detects Wi-Fi location and syncs recordings to a VDS for Gemini-assisted notes.

Scale / metrics

Under 2% daily battery use, approximately 18 MB per lecture, MQTT over TLS, SHA-256 checked atomic updates and 25 unit tests.

Architecture

[Termux Daemon] ⇄ Paho MQTT (TLS 8883) ⇄ Cloud VDS (Upload API + Gemini) ⇄ Android Intent Bridge [BBKClock + AAC + Wi-Fi]

Lessons & development

A Termux:API socket bridge handles restricted operations; the daemon handles Do Not Disturb and restarts after updates.

Technical detailsClose details

Problem

Investigating server failures from a phone requires log inspection, SSH and manual restarts.

Solution & process

A Telegram interface offers Gemini-assisted diagnosis, a session shell and coordinated restarts.

Scale / metrics

Supervises FunPay and PlayerOK services, restarts daemons in parallel and monitors load.

Architecture Telegram (aiogram) → Systemd & Process Supervisor → Log Tail Analyzer → Gemini AI

Lessons & development

Guarded remote shells, LLM-assisted DevOps and resilient daemons.

Technical detailsClose details

Problem

Basic local model CLIs lack tools, web search and persistent context.

Solution & process

A keyboard-driven TUI supports function calling, file reading, long-term memory and a Python sandbox.

Scale / metrics

Published as pip install w1x-cli, with open source on GitHub and Windows, macOS and Linux support.

Architecture Rich TUI / prompt_toolkit → Tool Engine → Ollama → Memory Storage

Lessons & development

TUI design, GPU and RAM profiling, and open-source distribution.

Technical detailsClose details

Problem

Manual Steam and Rockstar account delivery, Guard codes and slot tracking consumed seller time.

Solution & process

Automates delivery, IMAP codes, Steam Guard, Denuvo limits and slot reservation.

Scale / metrics

A 14,780-line plugin with session caching and transactional order storage.

Architecture Telegram / FunPay → Python Core → IMAP / Steam Guard → SQLite Storage

Lessons & development

The single-file design led to a modular architecture, coding agents and automated tests.

Technical detailsClose details

Problem

Manual order processing takes time and invisible input characters can break provider APIs.

Solution & process

Sanitises zero-width characters, integrates SocPanel API v2 and automates order fulfilment.

Scale / metrics

A 9,420-line open-source plugin supporting SMM panels and automatic cancellation on refunds.

Architecture Telegram / FunPay → AutoSMM Engine → SocPanel API v2 → Cardinal Delivery

Lessons & development

High-volume bot design, input handling and refund arbitration.

Technical detailsClose details

Problem

Manual plugin testing on live accounts was risky and made refactoring difficult.

Solution & process

An isolated setup runs seller and buyer, completes a test transaction, checks plugin behaviour and refunds.

Scale / metrics

Live end-to-end purchase, response and refund with logging and safeguards.

Architecture pytest → Isolated Process → FunPay Live API → Safety Guardrails

Lessons & development

Isolated verification, diagnostic artifacts and automatic restrictions for AI agents.

Technical detailsClose details

Problem

A lack of normalised seller and pricing data made market analysis difficult.

Solution & process

An asynchronous ETL pipeline rotates proxies, preserves raw MHTML with SHA-256 hashes, parses reviews and classifies listings.

Scale / metrics

2,000 sellers, approximately 108,000 reviews and 4,490 reference orders with a product taxonomy.

Architecture Asyncio Scraper → Proxy Pool → SHA-256 Raw Layer → SQLite Transactions

Lessons & development

Resilient ETL pipelines, data integrity, schema migrations and immutable raw data.

Technical detailsClose details

Problem

Plugin distribution and versioning require access control without exposing source code.

Solution & process

An update plugin connects to a FastAPI backend that compiles Cython binaries and a Telegram licence admin bot.

Scale / metrics

A central update hub builds for Python and OS versions, schedules module removal and manages licence plans.

Architecture Telegram Admin Bot → FastAPI (Uvicorn) → Cython Compiler → Hub Client

Lessons & development

Compiled distribution, update delivery and Telegram-based access management.

Technical detailsClose details

Problem

The goal was to present AI services through an interactive 3D experience.

Solution & process

A navigable 3D universe uses raycasting for selection, GSAP camera animation and adaptive rendering.

Scale / metrics

A public project with device-adaptive pixel ratio and optimised WebGL shaders.

Architecture Three.js (WebGL) → OrbitControls → Raycaster → GSAP → Adaptive Quality

Lessons & development

Learning a new technology, designing a 3D scene and delivering an interactive project.

03 / Tech Stack

Technologies

The stack behind these services, from low-level system calls to asynchronous backends and LLM pipelines.

Backend & Async Core

Asynchronous backends, queues, ETL pipelines and SQLite WAL.

Python 3.12+ Asyncio FastAPI Pydantic v2 SQLite (WAL) ETL Pipelines

Integrations & Protocols

MTProto, Bot API, MQTT over TLS and external web services.

Telegram Bot API Telethon MTProto FunPay Cardinal MQTT TLS aiohttp / REST API SocPanel API

Systems & Hardware

Systemd daemons, Android Termux, Windows services and ESP8266 firmware.

Linux C++ / ESP8266 C# / .NET Termux Cython (.pyd/.so) Win32 API

AI, Testing & DevTools

Static AST analysis, local models, pytest and terminal interfaces.

pytest Python AST (SAST) Local LLM (Ollama) Gemini AI API DAST Crawling Live E2E

04 / AI Workflow

How I work with AI agents

I use AI agents to develop and investigate code, set constraints, review changes and test the result.

Context and rules

I document architecture in PROJECT.md and working rules in AGENTS.md; agents start by inspecting code and constraints.

AGENTS.md PROJECT.md IDEAS.md Documentation

Tools for the task

I use skills for research, review and testing, and split well-defined work between agents where appropriate.

/study /critic /tdd-loop tg-bot-ui

Verify the result

I check findings against code, run tests, verify workflows and inspect services and logs for server changes.

Review Code review pytest Live E2E

Work together

Have an idea? Let’s discuss automation, a backend integration or an internal tool.

GitHub Telegram

W1bbex Python Backend Developer Automating manual workflows

About me

Projects and tools

PC Control

TG-FlowMap

Phone Control

VDS Control

W1x CLI

FPC-AutoOffline

FPC-AutoSMM

FPC-E2E

FunPay Parser

W1bbex Hub + Admin Bot

W1bbex AI Universe

Technologies

How I work with AI agents

Context and rules

Tools for the task

Verify the result

Work together