Eugene Svistunov

_VANS

Full-stack developer · DevSecOps · AI engineering

I design systems that survive production · Ukraine · remote

Open to projects

Telegram [email protected]GitHub

The full desktop lives on a computer. This is the short version.

About the developer

Experience

For the past two years I have been building enterprise-grade digital infrastructure — from the network perimeter and access policies to the services that live on top of it. In parallel I develop full-cycle web applications and bring the AI stack into real business processes: not demos, but systems that carry users every day.

Security and scale

Security and scalability are not a final phase — they are architectural requirements from the first commit. DevSecOps practice gave me the habit of designing access control, environment isolation, secret management, observability, and predictable releases before a project ever meets its first user.

How I cover the whole stack

I work as an AI-native engineer: architecture, review, security, and release quality stay with me, while implementation runs in pair with AI agents. That gives the throughput of a full team with a single person accountable for the result — which is exactly why I can cover the entire stack: backend, frontend, infrastructure, and firmware.

Projects

Web products, firmware, and server tools.

Web platform

Swish Swish

A platform for building and keeping wishlists.

It saves a gift from any store by its link, shapes them into a clean list, and shares it as a single link. The reservation mechanic lets guests coordinate with each other without spoiling the list for its owner.

Role: Product, architecture, backend, frontend, infrastructure — the full cycle.

Metadata import pipeline: a product from any store via one link
Gift reservation that never spoils the list for its owner
Token-based design system, server rendering without a heavy frontend build
A public link to the list — guests need no account

PythonFastAPISQLAlchemyAlembicJinja2 SSRES modulesDocker

swish-swish.com.ua

Telegram bot

Personal trainer bot

A coach's working tool, right inside the chat.

An exercise library with proper form guidance, individual training programs for every client, and online coaching: the trainer follows progress and adjusts the plan while the client receives workouts directly in Telegram.

Role: Built around the trainer's real workflow — from scenarios to release.

Exercise library with technique guidance
Individual training programs per client
Online coaching: the trainer tracks progress and adjusts the plan
Everything inside Telegram — the client needs no separate app

PythonaiogramTelegram Bot APIPostgreSQL

@imennotolik_bot

Open-source firmware

InkPoint X

A pet project that grew into a product.

Open reader firmware for the XTEINK X3 and X4 e-ink devices. Not a patch set but a complete system: library, file manager, multilingual typography, reading statistics, Wi-Fi book transfer, and OTA updates — all designed around a monochrome panel and physical buttons.

Role: Firmware architecture, interface, display drivers, releases, and OTA.

One interface adapting to two different panels and both button layouts
EPUB, FB2, PDF, TXT, Markdown, and XTC in a shared reader
Reading statistics, daily goals, and achievements
Over-the-air updates straight from device settings

GitHub stars: 73 ★ · 2 forks
Devices: XTEINK X3 · XTEINK X4
Panels: 528 × 792 · 480 × 800 mono
License: MIT

C / C++ESP32-C3PlatformIOUC8253 / UC8279DMIT

GitHub · dev branch Support on Ko-fi

Server management

Win XP Server

A Linux server with a Windows XP-inspired desktop.

A web interface for Linux server administration with a Luna desktop, Start menu, and movable windows. Manage real server files, use an interactive Bash terminal, monitor processes and resources, and open Docker web applications directly from the browser.

Role: Web interface and backend development.

File manager with uploads, ZIP archives, and a Recycle Bin
Interactive Bash sessions over WebSocket with support for terminal applications
Task Manager with CPU, memory, disk, and network monitoring
Docker container discovery and access to their web interfaces in desktop windows

PythonFlaskFlask-SocketIOJavaScriptxterm.jsLinuxDocker

GitHub

Skills

The stack is chosen for reliability and delivery speed, not for trends.

Backend — Logic, data, and integrations

PythonFastAPIaiogramTelegram Bot APISQLAlchemyPostgreSQLREST APIWebhooks

Frontend — Interfaces worth showing

TypeScriptReactNext.jsTailwind CSSSSRES modules

Infrastructure & DevSecOps — Security and scale from the first commit

DockerCI/CDCloudflareRailwaySecret managementEnvironment isolationMonitoringBackups & disaster recovery

Enterprise IT — The infrastructure a business runs on

Network perimeterAccess policiesEnterprise-grade infrastructureDevice management

AI — AI inside processes, not slide decks

OpenAI APIAI agents in the delivery pipelineProcess automationRAG & knowledge basesn8nSaaS integrations

Embedded — Code that runs on hardware

CESP32-C3 / S3PlatformIOE-ink and AMOLED panelsOTA updates

Services

Three directions where I take a task from idea to production.

Full-cycle web platforms

Product, architecture, backend, frontend, and deployment — one person accountable.

product platforms and internal tools
user accounts, admin panels, control dashboards
integrations with external services and APIs
server rendering and fast first load

Telegram bots & automation

Bots that close a real business process instead of just answering /start.

booking, support, and client-coaching bots
admin bots with roles and permissions
automation of routine work between services
CRM, calendar, and payment integrations

Infrastructure, security & AI

A perimeter where the product runs safely and scales without a rewrite.

infrastructure and access-policy design
secret management, environment isolation, monitoring
backups and a disaster-recovery plan
bringing the AI stack into existing processes

Solutions

Formats of systems I have built and can build around your process.

Online booking bot

Handles requests, manages availability, sends reminders, and syncs with calendars.

Who for: For service businesses with constant inbound bookings.
Outcome: Fewer lost requests and lower manual coordination load.

AI support assistant

Handles standard questions, pulls answers from the knowledge base, escalates complex cases.

Who for: For support and customer success teams.
Outcome: Faster response time and reduced pressure on human agents.

Internal AI knowledge base

Gives teams instant access to SOPs, docs, and internal instructions in one interface.

Who for: For companies with a large internal documentation scope.
Outcome: Faster onboarding and more consistent team decisions.

Lead qualification bot

Collects lead data, scores requests, and pushes prioritised records into CRM.

Who for: For sales teams that rely on speed and lead quality.
Outcome: Higher lead quality and better conversion of manager efforts.

Automation control dashboard

Centralised panel for workflow statuses, logs, failures, and manual checkpoints.

Who for: For operations teams running multiple automated processes.
Outcome: Clear operational visibility and fewer critical process failures.

AI product admin panel

Manage users, quotas, prompts, scenarios, and usage analytics in one control layer.

Who for: For owners of internal AI tools and SaaS products.
Outcome: Controlled product growth without operational chaos.

Process

So you always know the stage, the risk, and the outcome.

1DiscoveryWe map goals, bottlenecks, integrations, and the metrics the solution must move.
2Solution designI design the architecture, user scenarios, data map, and delivery plan before the first line of code.
3DevelopmentI build the backend, interfaces, and integrations so the system holds under load.
4Testing & launchWe validate the critical flows, ship to production, and set up control points.
5Support & iterationWe improve the system from real usage data and scale it as the team grows.

Principles

How I work, in short.

The business process comes first, the tool second — never the other way round.
Security and scale are designed into the architecture, not bolted on later.
Custom development instead of bending your process to no-code limits.
Direct communication with the engineer accountable for the result.
Production practice: validation, logging, monitoring, predictable releases.
Focus on automating routine work so the team spends time on decisions that matter.
Code and project structure stay fit for handover and further growth.

Contact

Describe the task — I'll reply with a solution format and the stages.