Roman Shtefanesa
Offer Roman work on your next project.
Private message
Offer work
Dnepr,
Ukraine
5 days 3 hours back
available for hire
on the service 3 years
Rating
Successful projects
Average rating
Language proficiency level
Русский: fluent
Українська: advanced
English: pre-intermediate
Skills and abilities
Programming
-
Bot Development
from 445 USD for month
-
C & C++
from 445 USD for month
- Cryptocurrency & Blockchain
-
Data Parsing
from 445 USD for month
- Databases & SQL
-
Desktop Apps
from 445 USD for month
-
Python
from 445 USD for month
- Web Programming
Services
Administration
-
Software & Server Configuration
from 445 USD for month
-
Windows
from 445 USD for month
Writing
Portfolio
-
Sim app
Web Programming
The backend part of a comprehensive system for managing telecommunications infrastructure has been developed. The project was carried out by a commercial team (backend developer + DevOps) for a telecom operator that needed automation for accounting SIM cards, ports, and numbers of physical and virtual devices.
Features important to the client:
- Asynchronous balance checking for 1000+ SIM cards using Celery and Redis, with an automatic retry mechanism in case of failures.
… - Intelligent time binding of equipment (e.g., devices to ports) with automatic restoration of original states after session completion.
- Modular Django admin panel with logical separation of functionality into blocks (SIM cards, ports, numbers, devices).
- Import and export of data in popular formats (CSV, Excel, JSON) for integration with billing and accounting systems.
- Detailed logging of all operations (who, when, what changed) for auditing and incident investigation.
Results:
- Reduction of manual infrastructure management time by ~70% due to automation of routine tasks.
- Full transparency of the status of SIM cards and equipment in real time.
- Ability to quickly generate reports on balances and bindings for accounting and technical support.
- The system has been successfully tested under a load of up to 50 simultaneous objects.
-
Comprehensive system for managing the results of psychological
Web Programming
A platform for automating work with the results of psychological tests (Cattell 16PF, Leongard-Schmiesek, Holland). The project is designed for psychologists, HR specialists, and researchers who need convenient collection, storage, and analysis of testing data.
Features:
- Automatic parsing of results from web pages using Selenium (support for three types of tests).
… - Unique slug generation for each test attempt – preventing data duplication.
- Relational database with abstract models and a constants table (minimizing duplication, optimizing storage).
- Django admin panel with inlines, search, filtering, and a custom form for launching parsing.
- REST API with filtering by user and test type – easy integration with other systems.
- Visualization of results (graphs, charts) and saving screenshots of profiles (Cattell test).
Results:
- Reduction of manual work in processing tests by ~80%.
- Centralized storage of testing history for all users.
- Quick access to data through the admin panel or API.
- Simple launch of parsing from the admin panel – data updates in a few clicks.
-
Inter Exchange Arbitrage Bot Architecture and Deployment
Cryptocurrency & Blockchain
A professional modular bot has been developed for the automatic search and analysis of inter-exchange arbitrage opportunities in the cryptocurrency spot market. The project was created as a personal initiative to hone skills in high-load asynchronous systems, integration with exchanges, and AI analytics.
Features important for the client:
- Automatic arbitrage search on Binance, Bybit, OKX, KuCoin, YoBit, taking into account fees, liquidity, and slippage (order book analysis).
… - Technical analysis (RSI, MACD, volumes) to increase signal accuracy.
- Aggregation of news from 7+ sources (CoinMarketCap, CryptoPanic, NewsAPI, etc.) with cascading failover between AI providers (Gemini, Groq, OpenRouter) — translation, sentiment, trading recommendation.
- Telegram bot on Aiogram 3: FSM, admin panel, customizable profit thresholds, P&L reports, order book density screener (Matplotlib graphs).
- Fault tolerance: Circuit Breaker for exchanges, rotation of HTTP/SOCKS5 proxies and Tor, distribution of API keys for news services.
- Full asynchronous operation: FastAPI + asyncpg + SQLAlchemy 2.0, background scheduler APScheduler.
Results:
- The bot consistently scans the market in real-time, processing up to 100 trading pairs per cycle.
- AI news analytics reduce false triggers on volatile events.
- The system has been successfully tested in the Bybit and Binance Testnet environments.
-
Fit Trainer Bot — Telegram bot for fitness and nutrition
Bot Development
Fit Trainer Bot is a Telegram bot that serves as a personal fitness trainer and nutritionist. The project was developed as a freelance task for a client who required an automated solution for online client management, including the generation of workout and nutrition plans, dietary analysis through photos, and payment processing for subscriptions.
Features important to the client:
- Full automation: the bot collects anthropometric data, goals, injuries, and equipment, and based on this, generates individual workout and nutrition plans (algorithmically + GPT-4o).
… - Food photo analysis through Vision API: dish recognition, calorie and macronutrient calculation, and tracking of food and water intake.
- Conducting workouts with a timer, prompts, and the ability to substitute exercises.
- Integration with the payment system YuKassa: paid subscriptions (basic/premium), webhooks, auto-renewal, and tariff recalculation.
- Web interface on Django for the administrator (user management, tariffs, log viewing). A health monitoring page for all services (DB, Redis, Celery, Telegram API).
- Fault tolerance: Docker containerization, automatic restart, and error handling during connection drops.
Results:
- The bot is fully functional and has been tested on real users.
- The client received a ready-to-launch solution with documentation and deployment instructions.
- Time savings on manual client management and increased retention due to personalization and reminders.
-
Solana Telegram Wallet
Bot Development
A Telegram bot has been developed for managing cryptocurrency wallets on the Solana network (testnet/devnet). The project was created to participate in the Solana Renaissance hackathon as part of a team of 2 developers and 1 organizer. The bot allows users to securely create new wallets, connect existing ones, check balances, transfer SOL, and view transaction history—all directly in Telegram.
Features to note:
- Full asynchronous architecture on aiogram 3 with a finite state machine (FSM).
… - Integration with the Solana blockchain via the solana-py library.
- Caching of balances and transaction history (TTL) to reduce load on RPC nodes.
- Validation of all input data (addresses, private keys, amounts) and error handling.
- Two database modes: SQLite (for simplicity) and PostgreSQL (for production).
- Web interface on Django for administrative monitoring (optional).
- Ready to run in Docker (container with PostgreSQL).
Results:
- The bot has been successfully tested on testnet and devnet, consistently handling requests.
- The repository contains complete documentation for installation and setup.
- Experience gained in working with blockchain, cryptography, and creating secure transactions.
- The project can be adapted for mainnet and used as a basis for a commercial crypto service.
Stack: Python 3.12, aiogram 3, solana-py, SQLAlchemy 2.0, Django, Docker, Git.