PAGE
SKILLS
My skills are better described not just through a list of technologies, but through the type of tasks I can solve.
Finding IT solutions
I work on tasks where you need to connect business logic with technical implementation. First it is important to understand the process: the input data, the people involved, the decision points, the constraints, the risks and the expected result. Only then do you choose the form of the solution — code, a spreadsheet, a bot, a website, a database, automation, an AI tool or a combination of several components.
Python and applied development
Scripts, parsers, data processing, integrations, working with APIs, files, spreadsheets and internal tools. Async web services with FastAPI (SQLAlchemy, Alembic, Pydantic) and server-side security: password hashing (argon2), server sessions, CSRF and SSRF protection.
Databases
My main database is PostgreSQL. I design tables to fit the task, store and process data, write SQL queries, and connect the database with Python scripts, bots, websites and other tools.
Spreadsheets and reports
Excel and Google Sheets: formulas, filters, large spreadsheets, data processing, import/export, automatic updates, Google Apps Script and VBA.
I use spreadsheets not just as a place to store things, but as a convenient interface for data, reports and workflows.
Telegram bots and internal tools
Bots for notifications, reports, data monitoring and interaction with internal systems — wherever a person needs to get information quickly, without constantly opening spreadsheets, admin panels and databases.
Web development
I can build websites, interfaces, portfolio pages, simple web apps and visually tidy solutions.
AI tools
ChatGPT, Cursor, Claude and others. I learn technologies faster, write and review code, find bugs, design solutions and generate ideas.
AI for me is not a "do everything" button, but a working amplifier that helps me move faster from idea to implementation.
Building AI/LLM services
I don't just use AI as an assistant — I also build products on top of large language models: provider-agnostic access to OpenAI and Anthropic Claude, function-calling and web search, structured JSON-mode responses, resilience to invalid answers, prompt engineering and token-cost tracking.